Transcripts For CSPAN3 House Hearing On Facial Recognition Technology Uses At Homeland Security... 20200212

breakfast this morning. and they will join us accordingly. the ranking member is one of them. the committee is meeting today to receive testimony from department of homeland security use of facial recognition and other biometrics. without objection the chairs authorized to declare the committee in recess. good morning the committee is meeting today to continue examining department of homeland security use of facial recognition technology the part one of this hearing july of last year after news that the department was expanding its use of facial recognition for varying purposes such as confirming the identity of travelers. as facial recognition technology has advanced, it has become the chosen form of biometric technology used by the government. i'm not opposed to the use of the technology as i recognize that it can be a valuable tool to homeland security and serve as a facilitation. but i remain deeply concerned about privacy and transparency and data security and the accuracy of this technology and want to ensure those concerns are addressed before it goes any further. last july along with other members of this committee share these concerns and left this room with more questions than answers. december 2019, the national institute for standards and technology published a report that confirmed age and gender and racial bias of facial recognition. this found that depending on the algorithms african-american and asian american faces were misidentified 100 times more than white faces, although they tout the match rate for this system is over 98 percent. it is my understanding, that they did not test their report, or the figures do not account for images could not be captured into a variety of packages such as lighting or skin tone, making the actual match rate significantly lower. these findings continue to suggest that some of this technology is not ready for primetime, and requires further testing for widespread report. this identifying a relatively small percentage of the traveling public could affect thousands of passengers and likely would have a disproportionate effect on certain individuals. this is unacceptable. data security also remains an important, concern. last year a contractor experienced a significant data breach, which included traveler images being stolen. we look forward to hearing more about the lessons we learned from this incident, and the steps that it takes to ensure that biometrics data is kept safe. transparency continues to be key the, american people deserve to know how the department is collecting facial recognition data, and whether the department is safeguarded their rights when deploying such technology. that is why we are here, seven months later, to conduct our oversight. i'm here that we have witnesses from cbp before us, to provide us with an update and answer our questions. we will also have testimony from dhs us office of civil rights and civil liberties, this offices is in charged with the protection as it relates to the department activities, no easy task especially these days. be assured that the committee will continue to hold the department accountable, for treating all americans and ensuring that our rights are protected. i look forward to a robust discussion, and i think the members for joining us today. >> i welcome our panel of witnesses, our first witness, mister jean wagner, currently serves as the deputy executive for the office and border protection. in his current role, he oversees nearly 30,000 federal employees and manages programs related to immigration, customs and commercial missions. and we previously served as immigrations and customs enforcement, he is the director of information, technology laboratory at the national institute of standards and technology. he oversees a research program that focuses on enter opera biloxi, security, usability and reliability of information systems. without objection, asked the witnesses full statement and we'll summarize each statement for a few minutes, beginning with mr. wagner. >> german thompson, ranking members of the committee thank you for the opportunity on behalf of u.s. customs and border protection. i look forward to the opportunity to discuss this with you, since we are using the defenders and the report, we are confident that these are cooperated. the 100 different algorithms, the highest performing algorithms that minimal levels of demographic rates. the report also highlights the operational valuables. such as gathering size, photo eight, photo quality, numbers of photos of each subject in the gallery. camera quality, lighting, human behavior factors. all influencing the accuracy of an algorithm. that is why cbp is deploying the technology to ensure we contain the highest levels, which remain in the 90% range. and one important note is that they did not discuss the operational construct, to mission the additional impact they may have, which is why we've entered into an mou, but as we roll out the bomb a check based entry, exit system we are creating a system but is facilitative. agility requirements are not new when taking an international flight. they require travelers to establish their identity but entering and the parting the united states. they employ biographical and biometrics base procedures to inspect travel documents presented by individuals. verifying the authenticity of the document and requiring the person presenting it. these are not new requirement. use a comparison technology simply automates a process that is often done manually today. the shortcomings of human, manual review and making facial comparisons are quite documented. humans are prone to fatigue, sometimes that bias as they may not even realize to include born race and gender bias, fingerprint biometrics have also documented gaps in the performance, the small percentage of people we see we cannot capture fingerprints from, and study said document this as well as well as demographic coalitions based on age. we are all well aware of the issues of common names and a biographical based system alone so no system is perfect. but since the united states put a digital photograph and to the electronic ship on a passport it, would seem to make prudent sense that the technology may be useful in determination of the rightful document holder. it's more difficult to forge or alter a legitimate passport, as security features are much more stronger. but still vulnerable to a vulnerable person that is real but belongs someone else. using facial comparison technology, we have identified 252 impostors to include people using 75 genuine u.s. travel documents. the privacy continues to be integral to our by metric measurement, we are compliant with the terms of 1974, a government act of thousand into the paperwork reduction act of 1995, and department of policies that govern the collection, use and maintenance of personally identifiable information. we recently published updates to the attendances in the privacy impact assessment covering this program. systems of record notices have been publishes on the databases to store the information met three times as representatives of the privacy advocacy committee as well discussions with oversight boards, the dhs privacy and integrity advisory committee. in november, we submitted to the office of management and budget a role making that would solicit updates and amendments to the federal regulations. one final note is that our private sector partners must agree to document specific business requirements, they are submitting to photographs in this process. these requirements must be delayed after they're submitted and may not be retained by the private stakeholder. after the devastating attacks, we as a nation ask, actually make sure this never happens again? as part of that answer, the 9/11 commission report recommended that dhs should compete a by metric entry exit screening system, and that it was an essential investment and national security. they're answering that call and carrying out the duties by continuing to strengthen its by metric efforts and verifying that people are who they say. thank you for the opportunity to appear today, i look forward to your questions. >> i now recognize him to summarize his statement. >> good morning. german thompson, ranking members, thank you for the opportunity to appear before you today to discuss the department of homeland security's use of facial recognition technology. the agencies has -- i would like to make three overarching points and my testimony. first, the officer of civil rights and civil liberties has been added continues to be engaged with the dhs operational components to ensure use a facial recognition technology is consistent with civil rights and liberties. law and policy. second, operators, researchers and policy makers must work together to prevent algorithms and the use of facial recognition technology. the technology can serve as an important tool to increase efficiency and effectiveness of the department, mission, as well as the facilitation of lawful travel but is vital these programs utilized acknowledging in a way that is safeguarding our constitutional rights and values. to achieve these, we influence the ages policies and programs throughout their life cycle, and engage with department departments in the development of new policies and programs to ensure the protection of civil rights and liberties are integrated into the foundation. monitors operational execution and provides feedback regarding the impact of policies and programs. and fourth and finally, we investigate complaints such as complaints including allegations of racial profiling or invisible bias. we >> recognize the risks of facial recognition algorithms, as raised by this committee used and facial recognition systems to identify and mitigate the invisible bias. we will continue to support the collaborative relationship between the national institute of technology, the science technology director and office up by metric, and opponents such as u.s. border protection. and carrying out its mission, we advice dhs components by participating an enterprise level groups working on recognition issues, and we directly engage with component, we are regularly engaging cbp on the implementation of face recognition technology, we advice on policy implementation, for individuals wearing religious headwear or with a religious objection with, and individuals that may have the injury. even for home may percent challenges may not be possible. as the facial recognition problems evolve, they'll be collaborating directly with cbp to address additional impacts. further, they will engage communities to vote to inform, and address critical concerns. for additionally, we will evaluate alleged violations of civil rights to further inform our policy, we understand the successful recognition technology requires ongoing oversight and quality assurance. additional validation, and a close relationship between the users and oversight offices. in this way, it can be developed to write work properly and without and permissible bias when it achieves additional operating capability and then continually droughts entire project life cycle. >> we will need to work with the operational components to ensure that policies and practices evolve, so that the human heart are also focused on responsible employment. working in a manner that prevents and permissible bias in activities. again, i thank you for the opportunity to appear before you, and i look forward to answering your questions. >> thank you for your testimony, and now i recognize doctor remain to satirize his statement. >> i'm chuck roman, the director of information technology, laboratory at the national institute of standards and technology. thank you for the opportunity to appear before you today to discuss the role in the standards and testing for racial technology. in the areas of biometrics, we have been working since the 1960s, biometric acknowledges provide and ability to identify behavioral characteristics. face recognition technology compares images to available images for verification or identification purposes. this has improved the accuracy usability, interoperability and consistency of identity management systems to ensure us interests are represented in the international arena. this research has provided state-of-the-art technology benchmarks and guidance and us government agencies that depend on that recognition technology. testing program provides technical guidance and scientific support for recommendations of utilization of the technology to various us government and law enforcement agencies including fbi, dhs, cbp, and i arpa. the interagency report 8280 released december 2019, quantify the accuracy of facial recognition algorithms defined by sex and age and race or country of birth through many - - search algorithms. and it distinguishes between false-negative and false positive in the errors are application dependent. this but from 99 developers using for collections of photographs with 18. 2 7 million images of eight. 94 million people. they came first debate of this e i will address one application. there are boss positive differentials are most larger than those related to false negatives and exist across many tested. they may present a security concern as they may allow access to impostors. other findings are that false positives are higher in women than in men, and higher and the elderly and the young compared to middle aged adults. regarding race, we measured higher false positive rates in asian and african american faces, relative to those of calculations. there are also false positive rates in native american, alaskan indian, and pacific islanders. these effects apply to most algorithms, including those developed in europe and the united states, however, a noticeable exception was for some algorithms developed and asian countries. there was no such dramatic difference and false positives and one to one matching between asian and caucasian faces, where the algorithms developed in asia. well the nist studies did not study cause and effect, one possible connection as a relation between an algorithm performance and the data used to train the algorithm itself. i will not comment on one too many search algorithms. again, the impact of errors is application dependent, false positives in one too many search are particularly important, because the consequences could include false accusations. for more soccer, those the study measured higher false positive rates in women, african americans, and african american women. however, the study found that some one too many algorithms gave a similar false positive rates across these specific demographics. one of the most are great for them still under this group. this underscores one overall message of the report. different algorithms perform differently. indeed, all of our reports note wide variations in accuracy across algorithms, and an important result from the study is that demographic effects are smaller with more accurate algorithm. we are proud of the positive impact we've had on the evolution of biometrics capabilities, with the extensive experience, both and laboratories and unsuccessful collaboration with the private sector and other government agencies, nist is pursuing the reserved necessary to employ inter operable, secure, and usable identity management systems. thank you for the opportunity to testify and i'd be happy to answer your questions. >> thank you. i think all of the witnesses for their testimony. i will remind each member that he or she will have five minutes. i now recognize myself for questions. >> we will start off with you. part of your report was next generation technology, and i understand that we will use to review existing technology. >> we are not certain of. that we intend to continue our investigations, the existence of the specific algorithms that we test, those algorithms are submitted to us by the vendors. we have no independent way to correlate whether those are the identical algorithms that are being used in the field. >> so, part of what you said is how the technology is deployed depends on the application of the technology, explain that a little bit more to the committee. >> our approach is that the significant thing to be cognizant of is the risk associated with the deployment and the studies that we do in help to inform policymakers, such as members of congress as well as operators of these technologies, about how to quantify those risks, at least for the algorithms themselves. the deployed systems have other characteristics associated with them that we do not test. we tests only the algorithms currently. and the second point is that risk that comes from the emirates, is part of a much larger risk management that the operators have to undertake. for example, access to critical infrastructures is different than access to a phone that you might have. the risks are different in those two cases. >> thank you. mr. wagner, can you share with the committee the extent that we go to protect the information collected in this process? >> the photographs that are taken by one of our stakeholders cameras are encrypted, transmitted securely to the cloud infrastructure, where the gallery is positioned that the pictures are template highest, which means that they are turned into some type of mathematical structure, and cannot be reversed engineered. they are matched up with the tampa times photos that we have pre-stage in the gallery. and the response goes back with a unique identifier. >> so the comment that two to 3% of people who are misidentified, what are they doing to try to get that to zero? >> it's not that they are misidentified, we just did not match them in the picture or gallery that we did have of them? so we should have max them, and it should be at zero, but that is where we look at the operational variables. the camera, the picture quality, the human behaviors when the photo was taken the lighting, those two types, other than the age of the photo, and what we have seen in the report, is the gallery size, they are comparing against a few thousand here, at most. and the numbers of voters we have from the particular individual can impact which one we match against and your match rate, and then the age of the photo. if you had your passport taken at age 20 and you are now 29, and your face has changed, we are going to struggle to match against that, which is compounded by poor lighting conditions or the person moving when the photo is taken, or a poor quality photo. >> listen to what you just heard. how have you all dealt with the positions about this technology? >> we have received one complaint of reference. however we have not seen a trend and that is when we opened an investigation in this matter. we are working on the policy side of the house, advising them directly but we also hear from the community, through our community engagement round tables and we have heard concerns in those forums about facial recognition technology, and those are concerns were using to form our advice. >> soaking in vied with the committee with where you have held those forums around the country? >> yes, absolutely, we have round tables in about 18 cities, and not to say that these have not been raised in every location, but certainly and some and we will continue to have those discussions in the future at future round tables. >> i'm not sure if you have information on this but last month iranian and lebanese nationals and individuals who travel to iran, and lebanon, most of whom were u.s. citizens were targeted, detained and subjected to prolonged questioning of up to 12 hours at the port of entry, i understand it indicates people were also questioned based on their religion. which is completely unacceptable. i understand we have admitted to enormous mistakes in this incident, if you know how this happened, and what are they doing to ensure that it never happens again? >> so there is no national directive or guidance that went out, other than because of the things taking place in iran, concerns about retaliation, we put our field managers on alert, to be more vigilant about current events that are happening, and work with your local state counterparts and really just be vigilant. there was some more prescriptive guidance that went out at a local level, which we are reviewing right now, because there is a lot of concerning things that we saw in the interpretation up that guidance, and the management and oversight as that weekend was unfolding and people were being referred in different expectant and governing, and there were concerns about the government or like thereof of what transpired so there is an internal investigation that we are conducting, and when we get the results up that, we will then proceed accordingly depending on what those results say. >> were you aware of that? >> yes, as mr. wagner, said we do have an open investigation in this matter,. are you ready? >> i yield to the ranking member for the opening statement. >> i'm sorry for being late, we just got back from the national prayer breakfast, thank you. after the tragic events of september 11th, the biometric systems are essential to our homeland security, following the recommendation, congress charged the creation of an automated biometrics entry and exit system, customs and border protection have already demonstrated the capability of biometrics to facilitate travel and enforce immigration laws, government and private sectors have made enormous strides in the accuracy, speed and deployment of the systems, biometric technologies of all kinds have seen improvements. these advances and facial recognition algorithms and particular transformational, the national institute of standards and technology is the leader and testing evaluations, but the team have done incredible work to help congress, dhs and industries understand the capability of currently available algorithms, but i'm concerned that some of my colleagues have already jumped to the misleading just hours after we released the 1200 pages of technical data, the majority tweeted that this shows facial recognition and even more unreliable and racially biased that we feared. if the majority had taken the time to read the full report before tweeting, they would have found that the real headline is determined that facial recognition is now being adapted as no statistically detectable race or gender bias. in other words, they define no statistical evidence a facial recognition algorithms that dhs is adopting contains racial bias. i hope that my colleagues will listen as he explains how the report proves that race or gender bias is undetectable in most accurate algorithms. the reality is that facial recognition technologies can improve existing process ease by reducing human error. these technologies cannot and will not replace the final judgment of the officers. concerns regarding privacy's are well-intentioned, but these concerns can be fully addressed and how by mitch systems are implemented, and i look forward to hearing the steps that we are taking to coordinate and predict privacy and civil rights of america, but as i have said before this is not a solution, and doing so ignores the critical facts it is not racially biased and does not violate the civil rights of americans, it is accurate, and it does protect the homeland, i appreciate calling the hearing it is important for congress to be fragile educated on this, issue i look forward to getting the facts and i yield back. >> i wish you had heard the testimony because there was some testimony contrary. >> i recognize the gentleman for his questions. >> my statement is wrong, anybody can jump at it. >> i would never tell congress they are wrong, -- >> we are one of the few people that won't do that. >> literally, to my understanding, there is no statistical evidence, that there is racial bias. is that an accurate statement? >> thank you for the question. in the highest performing algorithms for one too many matches, the highest performing algorithms we saw, undetectable. the bias, the democratic differentials that we were measuring. we say our undetectable. >> what do you mean by undetectable? >> what i mean by that is that in the testing that we undertook, there was no way to -- let me back up and say, the idea of having zero false positives is a big challenge. >> did you test any algorithms being used by dhs? >> we tested from them, but there's no way to determine this is what is being used. that would be something that we would have to attest to. from our perspective, the vendor provides us algorithms that are black boxes that we test, just the performance of the algorithms that is submitted to us by the vendor. >> mr. wagner, we are currently working to implement any other algorithms? >> any's e3 algorithms? >> yes, we're using an earlier version, and are testing the version that was tested, and the plan was to use it. >> who else can practice in the test? is it accurate to say that some are more sophisticated than others? >> yes. anyone can be around the country, but also from universities and some experimental systems as well. >> thank you very much, and let's get clear. you don't have an operational anyway in the country. you are testing it. that technology goes into paying next month? >> the algorithm will implement next month. the earlier version is operational now. >> but the one we are talking about is not? >> correct. you mention that african americans get misidentified. >> we don't see this to statistical level, or one too many algorithms. for the verification, we do see one to one algorithms, we do see evidence of demographic effects for african americans, for asians. >> the chair recognizes them for five minutes. >> thank you for clarifying that, because it was hard to understand from your testimony. just to be clear, doctor -- can you remind your name? >> so, in a certain segment they have higher rates of mistakes or african americans and agents. is that -- >> it's correct that most of the algorithms in the one too many that are submitted to exhibit those differentials. and some do not. i'm just trying to clarify. thank you so much for being here, i'm from michigan so we have a long history of needing our officers to protect us, and all of our bridges. so can you help me understand if this technology is being used in any way and our bridge crossings in the northern border? >> no. not at the bridge crossings. >> but at the airport. >> so, while i recognize it seems to be a small number of times of these programs where there have been more problems with african american women, and asian america, and walk with the process where, you are an average citizen, you're an african american woman, let's say we employ this technology and it shows a positive, right? just walk me through that process and how you would deal with that at the actual border for that actual citizen? >> he wouldn't to show your passport, which is what you do and the person would manually review. what if it did not match. >> and if they show the passport, but the technology still showed a match, what does that officer do in the situation? if the machine is eying one thing, and the passport is saying another? >> we would go on the basis of the document presenting, which photograph we have identified the with, or which identity we've identified the with. >> and the person across the border and go on, just asking for the average person to understand how this is being implemented. >> where matching people against that passport photo, we have an electronic copy of that passport database -- >> staff, you guys are being very disrespectful. >> we preassembled a gallery of those photographs, and that's why matthew against. when the officers screen, they will see the photograph which should be also what is printed on your passport, which should also be on that electronic chip. look at you and make sure you are the same person. if it does not match against that, will have to figure out why. >> and when you figure out why, is that individual allowed to progress -- we go to windsor to see a concert, we go to canada quite often in michigan. >> it could be as simple as saying okay, that you. i will figure it out later. >> what happens with the data? so let's say a woman has gone through her concert in canada, what happens to her data where it is flagged that she asked matched against someone who was done something wrong. what happens in the department to that information? >> if you're a u.s. citizen, the new photograph we take is discarded after 12 hours, there's no reason for us to keep the new photograph. there is a record of the transaction that you crossed the border. if there is some type of error, our analysts would look at that and corrected. if you have matched, which happens very often in a biographical section, name, data birth to the wrong person, even though your biographical match is identical to someone else, that is when we can also use a facial recognition to help us distinguish between people with the common names, and we can put note in the system to advise the officers or suppressed set information from appearing when we query here passport to next time. >> tell me how this technology, where you've been implementing it at different land borders, understand in the south, tell me what are the results? how many people have you identified in a positive way that needed to be identified. tell me some to six to demonstrate the value of these programs. >> it's 43.7 million people that we have run through it to date, at all the different locations, and found, outbound, cruise ships, land border pedestrians. we have caught 252 impostors, people with legitimate travel documents belonging to someone else, and 75 of those were u.s. travel documents. remember, u.s. travel documents, the only biometric we have is the digitized photo that the state department -- there's no fingerprint record, and i'm not advocating for one. but there is not what. they're so the only biologic we have on a u.s. travel document is the digitize photograph. that's a worldwide standard. that ship is allowed to be opened by any country participating in that scheme that can access to chip and pull off the digital photograph, and do some type of comparison to that. >> in my remaining time, tens of millions of people that you have used to have gone through this technology, tell me a little more about your stats how many positive stories. >> our matt rate is 73%, that means that we could not find that person in that preassembled gallery which means we did not match anything there. we just did not find a match of people traveling, it could be various environmental or operational reasons -- how >> many were false positives? >> i'm not aware of any but there may be a small handful that i'm not aware of. as we built this and tested we're just that seeing that. >> thank you. >> we recognize the gentleman from texas. >> the 9/11 commission recommended the use of biometrics for those entering and leaving the enunciates and i believe that technologies our friend in stopping terrorists and backed actors from entering this country. we see that time and time again. my question is that my understanding is the entry exit program that american citizens cannot out that program. is that correct? >> yes. >> there's no requirement that all americans have to be subjected to this. people >> have to establish their identity and went to be determined through manual review or by using the technology that they are a u.s. citizen they're excluded from the biometric tracking requirement. if they can opt out of having their fingerprints taken to make that determination. >> just like we use with global entry. most of my constituents love global entry. i got a clear program, as did mr. katko, associated with tsa so that he could put your fingerprints down and get to the head of the tsa pre-check line. this made it easier for the traveling public but also the great thing is it does not lie. biometrics -- it's you. it's hard to think that. the last congress we passed, by bill, the biometrics identification transnational migration program, otherwise known as bid map, i know this is a different program but passed overwhelmingly. 272. it is a different program under the obama administration that jay johnson and i talked a great deal about how we can use bit map to identify when people are coming into our hemisphere that may change their names multiple times along the route to get to the united states, get their facial recognition don't. their names to, but neither biometrics. this has been, in my judgment, a very successful program in keeping terrorists, human traffickers and bad actors out of this country. in fact, this program has enrolled over 155,000 persons of interest, and 460 known and suspected terrorist, including arresting violent criminals and rapists involved and transnational criminal organizations. so, mister wagner, can you comment on why that program is so valuable to the security of the united states? >> it's critically important, because as you mention, people do change their biographical tails. most of our watch lists searches are biographical based. but if we can identify people, especially people traveling by air that we have national security concerns about, and they're entering our hemisphere, settling in central or south america, we can work with our partners there and establish on a biometrics basis who that person is. so no matter what identity they show up in later and who were they when they first flew into that hemisphere? >> the travel documents can change and passports are stolen, and manufactured. but the biometrics don't lie. >> they steal documents, borrow documents and the ability to get a document that looks like you, if you can pass by the visual inspection of somebody glancing at it, yes. that's where the risk is. >> it's unfortunate, the senate did not pass this bill. they saw a lot of legislation from the chairman and i, and that is unfortunate. i would hope that we can pass this bill again this congress, and i do think we have to look at civil liberties as well. but this is entry exit, it applies primarily to americans who would want to opt in, and foreign nationals, and bid map applies to foreign nationals themselves. so, i want to thank the witnesses for the testimony, and thank you for having us here. i yield back. >> the chair recognizes the gentleman from new york. >> thank you very much, i think our expert witnesses who testified before us today. it's time that we face the facts. and regulated facial recognition is not an option. we can debate and disagree about the exact situations where we should permit the use of facial recognition, but we should also agree that there is no situation where they should be used without safeguards against protections for privacy. right now, in terms of regulation, facial recognition is still in the wilderness. facial recognition technologies are routinely miss identifying women and people of color. although there are some promising applications for facial recognition, these benefits do not outweigh the risks of automating discrimination. we have seen what happens when technology is widely deployed before congress can impose meaningful safeguards, so let us all look before we leap. >> we have observed issues with screenings at airports, for example, we've seen passengers and particularly darker skin passengers not able to be matched, due to poor lighting or other factors. this track how often they failed to capture photos of sufficient quality for matching? >> we track the number, we don't on all the cameras so it's difficult for us to track what an airline, how many pictures they may be taking before they submitted one to us for matching. because in the department they on them. so we are tracking how many pictures we receive, and what our match rates against them are. >> i was wondering about the quality, because if the photo quality is not good enough of the accuracy of the matching is irrelevant. do you track the numbers of photos that don't meet your standards? >> you say you have all these other partners taking photos and now you're dealing with something that has become irrelevant. if you don't know what subset of those don't meet your quality control. >> well, we know the pictures that they transmit to us, whether they mean, and you don't know the quality? you don't know how much of that, what percentage of that doesn't meet your standard? >> >> you know the percentage? >> how does they plan to address these issues to ensure it can capture high quality images of travelers for successful facial recognition and screenings? >> that's the partnership with nist. we look at the operational variables to make that even more high performance. >> what i would say to you is that, then, until you have met that standard, you are not doing the public a service of quality control. >> we are developing that standard. >> right, it's not developed. you are developing. >> no, what we are seeing, for matching at a 97% rate, -- >> let me go to another question. it doesn't matter. >> we're not seeing demographic based rates in that 3%, and that's where we partner with nist, to help us understand that. >> i understand that they completed operational testing on the biometrics and the results indicated that the accurately maxed images, but the capture was significantly lower than expected. 80% compared to 97%. >> most of these issues were attributed to airlines reverting to manually processing passengers, to speed the boarding process, are you aware of these findings? >> yes, and that is as we were developing the operational valuables, can we make it work? now we can look at and work with the airlines to not shut down their boarding but is the ease of the application of the traveler engaging with that? >> what steps is cbp working with? >> we would then put the requirement on to the foreign national who has to comply with the biometrics congressional mandate then we can work with the carriers to increase the rate -- >> can you provide our committee with those steps? that me ask you because i'm running out of time. you spoke about impermissible bias, and i was wondering since use that terminology, is the inverse -- is that a part -- is there something permissible bias? >> if i understand your question, the reason we use that term impermissible bias is because there are lots of reasons why there may be a flagrant act like lighting, environment, but we are focused on an air that is created based on a particular characteristic like race or sex or age. when i make that reference to impermissible by is that's what i referred to. >> there's no bias that is permissible? in other words, if there is a quirk of some sort and you find it to be so inconsequential that it becomes part of your standard, that becomes permissible bias? i'm trying to understand what you mean by impermissible bias. >> one of focusing on is what is prohibited by law that our office would look at, which is based on those protected characteristics. now, of course, obviously folks across the department are trying to eliminate any bias or any reason, but however, in terms of what we do as a policy office, we are really focused on the potential for bias in those projected areas. >> i yield back, thank you, >> the chair recognizes the gentleman from new york. >> thank you, and i appreciate you having this hearing. this is very important. and i commend all my colleagues for the probing questions because it's important. but i would like to make a general observation. when i was first a prosecutor, dna evidence was this weird science thing that nobody really knew about. as we went on, and as you got refined, and as it got better, it became a very potent tool, not just for law enforcement, but to exonerate people who are wrongly accused of crimes and i see the biometric technology filling a similar role. it's going to help law enforcement, it's also going to do a dramatically good thing to prevent misidentification of criminal conduct and i'm heartened about. that that the highest performing algorithms have no statistical anomalies. so that means that at some point those algorithms will get to the front lines and i encourage you to get them to the front lines quickly and i would encourage you to never let your guard down and follow any problems with the systems to make better because we are all going to benefit. i trust my colleagues, so i have to ask something that occurred yesterday that is very important to my constituents. a lot of resent on february 5th, which was yesterday to new york state. saying why almond security can no longer have new york drivers license as part of the formula or the trusted traveler program. that is because the new york state, under the green light act permits access to the new york driver database. so can you summarize for us. i asked this to be incorporated into the record. >> without objection. >> can you summarize the contents of that letter? >> new york state, because of the law they passed, without consultation, shut off the access to motor vehicle data, which included drivers license information, license plate registration, vehicle registration information, so in our operations any up the work that we do where we would use that information to help validate identity and address in a vehicle, ownership of a vehicle is impacted by not being able to do that directly. and you know, the breadth of our mission goes way beyond, i think, with the law says about immigration enforcement and impacting the customs mission in the national security mission and all the other areas in which we operate. >> and is there any other state in the country that is having this problem with customs? >> we've worked some other agreements with other states to continue to access the data for the work that we do. >> so my to understand that new york state is only gone from its customs and border protection as well as ice to have driver databases? >> yes, with i'm familiar with right now. >> even in california? >> california we have a separate agreement with what we continue to access the information. >> i just want to know further, as long as we have a couple of minutes, here are some things that are in the letter. tell me if this is correct. on a daily, basis i.c.e. uses new york data to combat national gangs, narcotics smuggling, human smuggling and trafficking, trafficking of weapons and other contraband, child exploitation, child exploitation, exploitation of sensitive technology, fraud and identity theft and is it fair to say that might not having access to this database hampers investigation at times? >> any law enforcement past this where we would use that information would be impacted, yes. i >> yield back the balance of my time. >> thank, you the chair recognizes the gentlelady from new york, miss rice, for five minutes. >> thank you, let's continue, mr. wagner, if we can, talking about what what happened with new york. would cbp be aware of the policy before the acting secretary's announcement on fox news? >> yes. >> so you were aware of it. no notification was made to congress about blocking access to these programs for new yorkers. >> i don't know. >> well, there was none. so personally, we my offices already received an influx of new questions about this policy, literally overnight and a 50 to 80,000 new york state -- are affected who have pending global entry we newman -- applications or renewals. this will have an enormous impact on people, many of whom entered into this program because their jobs require them to travel internationally. so, what do you plan to do about all those people who are going to be impacted? >> well, without the ability to help validate their identity through -- >> you have the fingerprints. >> yes, but if they have been arrested, the fingerprints to tell us anything. what would fingerprints tell you if you haven't been arrested? >> somebody trying to find out, is my point? >> i'm trying to validate their address, where they live, the residency, these are things important to us as we established that low risk trusted traveler status that we have four people in that program. without the ability to do, that how would we do that? so if new york state would shut off of the consultation are access to that information in december, how would we continue to operate and validate who people are? >> well, going forward, but what the people who already have it. i have global entry. so when i go to renew it, i'm not gonna be able to do that. >> correct. >> and yet here i am, a sitting congresswoman with global entry, so, to me, to me, to me, i understand the distinction that you're making, there are 15 other states, you're saying you have individual remiss with all of them but they do not block access to this database? 50 other states -- >> i'm not aware of any other state blocking our access to that information. >> okay, so i would like, you are going to follow up, i'm developed erectly with you, because there are at least 15 other states that allow undocumented people to obtain drivers licenses. >> i'm not aware of them blocking our information. >> all right, so you're not being aware is not a sufficient answer because there could be other states that do, and it seems to me that this is once again an attempt by this administration, specifically donald trump who formerly was a new yorker, to punish new york, right? so you and i are going to follow up on this, and appreciate you trying to answer these questions but we need more information, right? i appreciate your attempt to answer those questions and i yield back, thank, you mister chair. >> thank you very much, mister wagner, just for the record, can a person have global entry without a drivers license? >> yes, i believe so. >> so i'm trying to figure out how are you going to cancel all of these people as some of them don't even drive, and deny them -- >> well, is a new york state identification. >> they have passport. >> validation of that information. >> but they have a passport. they have a passport. >> how do we validate the address of where they live? >> my drivers license has a post office box. so i'm just try to figure out, are you being -- >> why is the information block to this purpose that? >> i'm using, why would you cancel it? >> i would new york state block the information for this purpose? >> is it for identification of security? >> both. >> but you can prove it without the documents, i mean that's what i'm trying to, well, chair recognizes the gentleman from louisiana, mr. higgins. >> thank, you mister chairman. i yield one mill it to my colleague, mr. katko. >> thank, you mister higgins, and i want to quickly follow up with my colleague, first of all, it is clear that it hampers investigations with ice. it clear that it hampers are a -- are billeted to get certain information available from drivers and databases a new york state. just to make sure it's clear, my colleague from new york has raised -- and there are many other states that have possible allowing -- the issue is is there any other state united states of america that completely blocks customs and border patrol and i.c.e. access to dmv records? >> i don't believe so. >> so, in my opinion, and have immense amount of respect to my colleague from new york, i do not believe this is a political exercise. all of new york would have to do is enter into a similar agreement that those other 15 states have entered into with customs and border protection and i.c.e. where they simply verify that they won't use it for immigration enforcement purposes, they use it for law enforcement purposes and for global entry and those types of things, is that correct? you can do that? >> i think that's a discussion we would have with the state. >> okay, and you've done it with other states? >> yes. >> okay, thank, you i yield back. >> i think my colleague, and just a follow-up on, the new york question, because it's a fascinating topic, are you aware of negotiations of communications player to the new york legislative body passing this law with the customs border protection? where we out front in this communication at all? our access -- >> seems to me like i should know before they pass the law this was going to happen. >> right, our axis was just turned off one day in december and our officers, agents in the field called and said, what happened to our access? >> so you're saying that as far as you know, and you can certainly advise if you do not know or have no way of knowing, but as far as you know, sir, was there any ongoing communication towards the development of this legislation in the state of new york with law enforcement agencies like customs and border protection and i see? well, one would hope that there was. doctor roman, you mentioned blond box testing. what you clarify that means that as your products have tested, facial recognition products provided, they are tested without your knowledge of who the vendor is? you're strictly looking at the results of the algorithm themselves? >> when i use the phrase black box testing, we don't have any insight into the characteristics themselves, they -- you >> know the identity? >> it's self identified. >> itself identified, as you study them. >> can any vendor cement and algorithm to this? >> yes. and the process by is submitting that product is standardized? >> it is. the top performing algorithms like customs uses, is there a wide variance between what your performing two and academic projects, perhaps submitted. >> there's a wide variance -- in your scientific assessment of missed testing would you say that what you're referring to as the top performing algorithm has been used by customs and border protection are far and beyond the common projects? >> the top performing algorithms are significantly better -- >> can you confirm for this committee that it is the top performing algorithms at this point, as being used by our federal on foreign agencies? >> i have no way to verify that. >> we just said that customs and border protection has been used. >> can you confirm that? >> not the algorithm we tested but the previous version of it, or upgrading to the version they tested last month. so -- >> going to the next iphone? >> i think that vaguely answers my question, and my time has expired. >> the chair recognizes the gentleman from new jersey. >> thank you. . we who and where is this facial recognition data stored, and please describe, under what specific circumstances this is allowed to be shared or used and transferred if that is the case? >> we are using, as a database, travel document databases. so these are photographs collected by the u.s. government for the purposes of putting on a travel document, and a u.s. visa. or a photograph that a foreign national, when they arrive we would take their photographs, or read the photograph from the chip in their passport and store that. so that's what forms the baseline gallery that we match against. new photographs we take of a person, a u.s. citizen -- or u.s. identity, those photos are discarded. after 12 hours, just for some work. if you're a foreign national, that goes where they are stored under the protocols of the record notice of the data retention period, which is 75 years. >> to follow up, we are living in an age where everything is being hacked. what part of a measure has been put in place? >> the data pinned is used in the u.s. government, cbp does not necessarily keep your own any of those permanent databases and are owned by the department of state and other branches of dhs, we use it, and then we use information back into them. >> i continue to have hits come across my desk about the mishaps and disadvantages of facial recognition technology and the racial bias. the technology continues to misrepresent and identify people of color and women. so am i hearing from the majority of the panel that that is not the case? it keeps coming to us, so it has to be some validity. >> and our testing for the one to one identification algorithms, we do see evidence of the demographic effects, differences with regard to race and sex and age in that one too many identification testing that we did for the algorithms that we tested, there is a small set of high performing algorithms that had undetectable deferential. but the majority of the algorithms still exhibit those characteristics. >> can you give a description of the two? >> in the case against verifying and identity, a biometrics is matched solely in the case -- >> is that the one to one? >> that's the one to one. to try to determine if you are who you say you are. and it's matched against a gallery of one, in essence. >> the one to one is in cbp's application, one to thousands for the airline and won two millions in the case of law enforcement, such as fbi to try to identify the suspension. >> so you're saying the percentage, of identifications in the one to one, you have more incidents of this bias? >> i should clarify, in the algorithms that we tested, that is correct. however, many of the vendors who chose to participate in the one to many testing did not choose to participate in the one to one on those are some of the highest performing in the one too many. >> thank, you and i yield, back. chairman >> thank, you the chair recognizes the gentleman from north carolina. >> thank, you mister chairman. i'd like to yield one minute to the gentleman from louisiana. >> thank, you my colleague, had a question regarding the fact -- effectiveness of this technology that you test, regarding children. is it a potential -- if we assemble the gallery of photographs of children, some of whom were being exploited, and fall certifications were presented, how does the technology work with children compared to -- mistakes and errors and other demographics? and can the second algae be used to protect children that are perhaps being exploited in crossing our borders, coming into our country, and if, so what can we do to protect the privacy of those children, given the fact there minors, and i'll give you moderating 30 seconds? >> thank, you sir, the application specifically is something that we don't test. what we, do what we have tested is the effectiveness of the algorithm in terms of error rate and we do find that four children in the one to one setting, the one that you just described, there are demographic effects, there are differentials, the error rates are higher in the one to one case, with respect to age, so it's more difficult, based on our testing, it appears more difficult. >> but there's no gallery. there's no one too many. >> we have no such gallery. >> if you did develop, that this could test the effectiveness and perhaps this could be a tool to protect children? >> we could undertake many different kinds of testing to determine the effectiveness of this. >> thank you sir, i thank my colleague for yielding. >> thank, you representative higgins. mr. wagner, is it true that by a metric entry exit system uses less personally identifiable information than the current system that we have in place? >> yes, because currently you open your passport book, let you show it to an individual to either say, check your bags, go through he has a screening, you're exposing your name, your date of birth, your passport, number all the information on your passport. somebody could be looking over your shoulders, you can take a picture over your shoulder looking at, that you're disclosing it to a person who doesn't actually need to know all of that additional information, versus standing in front of a camera with no identifiable information other than your face, which they can already see, and your picture is taken on the screen comes a green check mark, not person i know you've been validated by the government record, supersede so you're sharing actually last information in this instance. >> not only showing this information bra scale of one to 10:10 being the highest, how would you rate this progress has, in your own, words has continued to develop, and rightfully, so would be the highest security possible for travelers compared anything else we're doing now? >> right, now on top of everything else we're doing it brings us closest to ten, which is where we want to be. >> when representative higgins talked about some of the children involved, either a numbers for statistics based on people you have caught either involved in human trafficking or some other nefarious activity caused strictly because of the facial recognition? >> yes, on the land border we've got 247 impostors so far, meaning they had a legitimate document that belong to somebody else. 18 of those, so, 7%, were under the age of 18, and would be considered children. 73 of those at the land border had u.s. passports or u.s. passport cards and 46 of them, or almost 20%, had criminal records they were trying to hide. >> do you believe that these were identified strictly because of the use of facial recognition or was there any other aspect of their involvement? >> our officers are also very good at identifying behaviors in the person when they present their travel document, a lot of times that can also be a cue that the person is hiding something, but the technology on top of the officers skills and abilities should bring us to that security posture, that will bring us to new york for effect. >> are there any policies difference between a u.s. citizen versus non citizen? >> everyone has to establish their identity by law. everyone has to produce some type of identification. the law requires you to turn for the passport. >> the process of forming this 12 hours after is what? >> that's our return policy, we take a picture, discarded after 12 hours, looking at actually shrinking that to last time. we only keep the my case is what we have to restore everything. >> thank, you i yield remainder my. time >> chair recognizes the member from las vegas. >> thank, you mister chairman. i i find this interesting, the more you talk, the less i know. my karen airport is in my district. it's a very busy airport, one of the busiest in the country. a lot of international tourists come through there so i know we've talked a lot about the use of this facial recognition for security reasons. at like to talk about it in terms of how it affects the passengers experience. we want people in las vegas to have a good experience from the time they land the time they leave, so, how do you work to coordinate using this for security and also reducing wait times, or serving the passenger, as opposed to making it more difficult for the passenger? mr. wagner? >> so, it absolutely supports our travel and tourism goals as well. it makes a must-see better passenger experience, a more convenient passenger experience, a more consistent passenger experience. because you go through the airport, never stops you have to make to produce a piece of paper or open your passport again or provide some other form of validation to go forward, you can use the facial recognition and the camera to have the same process and it's quick enough that you walk up to get your picture taken, into the three seconds you moving forward. reducing the irthe airline incorporated into the voting process, or reducing their boarding times of the aircraft, sometimes as much as, say 40, 45%. it's a different atmosphere for the traveler, because you're not fumbling for your documents or getting stuck in line behind the person who's phone or forgot where they put their passport so it's creating a better atmosphere for the traveler. it's moving the lines quicker because it's creating, you can't leave your face on the plane, you can't leave your face in the bathroom, you can't forget that like people do with their travel documents so it's making an easier process because everybody knows how to take a picture and what we see is people are enjoying this process a lot better and what we're seeing is, the line is reduced. >> are you working with tsa or local law enforcement to make this all run smoothly? >> so, where we're coming very closely with tsa. we have an ongoing pilot in atlanta where because we built the gallery, as the person prints out there boarding pass, any place now where they have to show their passport at the airport say, when departing the u.s., you could take a picture and validated against the galleries so you're outside of the airport or you just walked in, you've got a boarding pass, the picture goes into that gallery, so, steps like checking your bags, where you have to show your id to the airline, person, you can have a camera there. gallery again because we built it for the biometric exit requirement. we would want to make a environment available to all of the other places in the airport where he would show a passport to do that, so yes for tsa you could take a picture and go through screening. you go to board the plane into the airline takeandthe airline e and comes back to the gallery, we confirm it and you can board the plane without even showing your passport to the airline or your boarding pass. >> suppose they find somebody doesn't match. my understanding is this goes through an application and law enforcement if they are busy or a person is responsible for checking things out, you have some kind of staffing model backs who is responsible for that because i haven't heard them since there's no action you are supposed to take that very clear, sometimes they just ignore it. >> depending on where you are, generally it would be the airline or the cbp you advocate the physical passport which is what you are presenting no now f we would make that determ and we would make a determination now,. we may ask a person for another form of id or additional question. we may do a further inspection on them so if you don't look thk like your passport photo from the visual, these are the same kind of things that would occur. >> we have been having a lot of confusion about going to the real id from the just regular driver's licenses. people don't know they have to do that. we are trying to get the word out. some states don't provide the funding. is that the transition part of your consideration as you develop? >> people will use their passport. >> these are international travelers we are talking about today. >> you don't see this moving to the national as well as international requirements of how this might apply to the domestic light i think there's some good discussion to have and should the traveler opt into this would bin tothis would be a system if that is what people would want to. >> the chair recognizes the gentle lady from illinois. >> the gentleman from texas, mr. crenshaw. >> thank you for being here it's been an interesting hearing to watch. it seems to be abnormally controversial. it's not like china or facial recognition or the time square downtown, talking about facial recognition, air land and sea ports where they have the duty and responsibility to know where they are already checking for identification of course. it seems from the answers we've gotten that it's using the best algorithm and that is what we've established today as far as i understand. the locations are marked. >> it is where you would normally present your passport. >> entrants are allowed to opt out of the facial recognition technology and provide identification to be a officer to be presented. >> it is stored for no more than 12 hours and encrypted virtual private cloud. >> the biometric data but are the major privacy concerns and how can we improve this? >> what we have heard from the community is people get used to the convenience of this technology and that bleeds over into the commercial world and they would be more likely to allow that to happen outside your government requirements. my discussions with them i said but there's also an explication by the public when they have this convenience in their private life, why should interactions be so integrated? why should the travel through the airport be so antiquated and frustrating? shouldn't the same apply when they are traveling internationally? >> one way this can be used to combat human trafficking, is there a way that tools like this can be integrated with others like spotlight, safer to battle child sex trafficking in human sex trafficking? >> what this helps us do in the core vetting process it by a graphically based. the name and date of birth is submitted to a watchlist through airline applications were tsa. we do those background checks on the basis of who the airlines tell us is lying and who purchased the ticket. when you can use a biometric to validate that you've added the right person, you have the assurances that that is a person that is actually traveling and not just their passport under a different person that's being trafficked. so it helps us close the vulnerabilities of imposters for nefarious or being trafficked were being victimized to be able to do that using the imposter documents. >> in my limited time left, can this be used to combat visa overstays as well? >> we track the visa overstays primarily for the biographical information the airline provides, but by implementing the system, we've actually biometric reconfirmed almost but i implementing the system we've actually biometric we confirmed almost 44,000 them over stays with the biometrics validation that these people overstay and leave the u.s., just about 44,000. >> thank, you i yield back. >> the chair recognizes the gentleman from new york. >> thank, you mister chairman. mr. mina, the nypd has in the past used facial recognition to compare photos from crime scenes against its own internal arrest databases. some state lawmakers want to take that ability away from the nypd and other new york state law enforcement agencies. do you support police agencies using facial recognition in the course of their -- have focused less so on the identification piece and we are looking at a much narrower - - narrower if i understand technology correctly per and whether that is based on race or natural - - national origin. >> the way that civil liberties is absent for the use of technologies. >> and with the false positive for those who are being arrest and making a false positives. i'm thinking of people who are being over policed, questioned further based off of just a verbal description. >> absolutely, congressman. i think it's obviously critically important to blend both the use of technology as well as the end user in this process. i don't think it's an either or proposition and as we have advised cbp, that is from a policymaking perspective, that is really where we see the greatest benefit, it is really that interaction with the technology, as mr. wrecker talked about earlier. for example and if it matches. >> and so we are all on the same page the use of technology has consistently preserve public safety and also to further protect civil liberties. and that we are unnecessarily politicizing. and we have some work to do to make it better. >> i am not a supporter of the new york legislation passed. i think it is unfortunate and wrong that you were not notifie notified. but two wrongs do not make a right. i have very simple questions and to do this professionally do you so i'm asking you all some very simple questions. if you are sitting out to be the professional force that you are into this professionally, do you think the in advance of announcing this, you should've told congresshs. >> this is ridiculous and it's a simple question it is a simple question. we heard about this from fox news. this is politics at its worst. we are talking about acting my professionals right now. if there is a problem that needs to be addressed we are talking back in the professionals right now. if there is a problem that needs to be addressed and you all are doing this, do you think it was appropriate that we were not told well in advance so we could try to arrive at some solution? do you think that is okay? is that the way you would want to carry this out? >> i'm not going to comment on that. all assume what you were thinking and unwilling to say right now. >> members of congress who cares about them there's millions of other people my staff or colleagues are all types of people it's politics. if you are making an effort to address a problem it would've been a system or proposal a conversation, that is the way business is conducted. so let's put that aside. so where do you now commit that we are all engaged in politics. conversations on the way fair lawn - - on the way forward. >>. >> thank you. >> many of my constituents with a major airport in chicago we are always interested for those that can improve security wait time. and then to use facial recognition that is fair and reliable in the effective way. and they go through that collects mr., wagner although childrennie this information? >> if you are outside the scope 14 through 79 i believe they discard all that information. and with this policy. >> are there any with the children biometric information. >> and then to find that children are more likely to be misidentified and that they are also misidentified. in taking the patterns to correct in this report. >> that we are not seeing those demographic based errors. now if someone does not match to the gallery or the document they are presenting. it with that confidence it is the person we can do that through questioning or addition identification. we could do that through an inspection of the person. sometimes it is just looking at the passport and saying, go ahead. it all depends on how discrepancy you look from your travel document photograph. >> and some passengers report being unaware or confused about how to opt out of this screening. as cpp expanded biometric screening program, does it tend to re-evaluate the best method of communicating the information to passengers? >> yes, sir right now we have signage at the airports, a lot of people don't read signs at the airport. we have gait announcements that the airlines try to make before boarding. again, as i was competing announcements going on and sometimes it's tough to understand what's being said so were actually looking with the airlines,, can we put things on the boarding pass, could given notifications when people are getting their ticket or their check-in information for boarding on their electronic messages, so we are looking for additional ways to do that. we also started taking out some privacy advertisements advising people of the requirements and what their options are as well. >> all right, well, it is certainly my interest of making sure that every passenger understands that one, this is, happening into, that they have a choice to opt out and i would certainly urge cbp to strongly consider, and issued this committee a timeline for perhaps outlining how we can improve that communication to all passengers. the cbc -- cbp after the rate of pulse -- false positives or mistaken identifications or the technology is used? >> but we track are the people where we take a photograph and are not able to match it to their travel document in our gallery and that is that two to 3%. i'll review of that information does not show discrepancies. >> that was not my question. my cautious captioning and reporting by put event. we want to know, the false positives, are we seeing more certain places along the border, or? >> you're not seeing false positives. that's matching you do a different identity. were not seeing that with this technology. >> or mistaken identities. >> we are not seeing that. more likely, you don't match against anything so we get no information returned. >> okay. doctor, can you list the recommendations to improve accuracy across demographics? >> the report, the testing that we do does not result in recommendations specifically to the vendors other than to take the dated that we provide the evaluation results and strive to use those results to improve their methods. >> so you're saying you don't have a lot of interaction with the developers? >> we have informal interaction with them in the sense that we have the scientists to do this by metric testing are part of a larger biometrics community, the vendor representatives, the scientists and so on but with regard to the testing, the feedback that we provide to the vendors is the digital. >> okay, so you all are doing like meetings with industry and helping them improve the quality of their product. >> we do host events but more as a convenor to get the community together to discuss different techniques. we don't provide, other than sort of in the general scientific community sense, we don't provide specific recommendations. >> okay, i recognize my time's expired. we just need to get more information about that and writing, will send your question. thank you, sir. >> thank you very much. the chair recognizes gentlelady from new jersey. >> thank, you mister chairman, and thank you for your testimony. a couple of questions. i think i just want to talk more about the role -- it seems to me there has not been much coordination across the dhs spectrum on directions from the adjust each component regarding the deployment, and the biometric in all, jay-z can correct me if i'm wrong. is there any sort of department line strategy in place for the use of biometric technologies or are components like yours given wide latitude to stand up biometric programs as you please? >> i'm sorry,? >> are you a lone ranger? >> are we what? i'm sorry, i didn't hear that. >> this coordination, it does not seem like there is the sort of department wide oversight. i want to know whether or not you are getting directions from others, because then i want to ask mr. miller, what is your role, and to what degree have you been involved in the oversight and in signing off on how these things can be done? >> all right, so, as we build a new programs, we are bound by certain statutes that require us to publish say, your systems of notice, your privacy impact assessment, where things are reviewed by our internal counsel or privacy officer and make sure we meet all the requirements of the statute, do we have the authority to do what we're doing? it is our timeline for storing and sharing it,. >> so, are you operating within your silo, this is what the law says with regards to what you can do, is this how you execute based on what your interpretation is of that, or is there a dhs component that plays into this as well and says, okay, but this is how we want to see this? >> depending on the acquisition process, there is a multitude of people at dhs that look at the acquisition, the resources spent, there is a whole process to go through for approval, at various ports, that authorize the expenditures and the investment in that. there is the dhs privacy, officer dhs council so a lot of oversight by dhs already in this process. certainly the rule makings we go through dhs council, the dhs policy, so there's a lot of oversight in coordination. >> it is my understanding, though, that there is no centralized body that gives the program a stamp of approval or certifies they're ready for prime time. is that correct? the crcc l approve your program? do you know? >> no, that would not necessarily go to them for approval. >> well, approval in the sense of maintaining objective privacy right. >> things are reviewed by them and i will defer to my colleagues. >> what did i answer that nicola different ways, congresswoman, to let me step back a second, and talk about the first part of your question, regarding the enterprise level review? i think one of the many ways crcc all participates in that dialog is by serving on enterprise level wide working groups across the department that include representatives from cbp, dhs, and he, and the office of biometric identity management where we actually are talking about a lot of these issues and we don't have a privacy impact assessment type model, however, we do work very closely with the privacy office regarding, not just facial recognition technology but certainly other forms of biometric edification. with regard to our list to put cbp we work with them in a couple of different ways. first is very directly, in terms of operating advice and then also, we go on site and work with cbp and privacy and civil liberties are important there as well. >> so, is your role anything more than just advice, observation advice? you have no authority to say, no, that's not working, that's a violation? no? and vice? >> that's not entirely accurate. but i would say is yes, we do have an advisory capacity. we have a compliance function where we do offer recommendations to components based on investigations. >> and if they don't all of them? >> i one last question. it has to do with the whole system that is used when were taking pictures, who is in charge of determining whether or not the lighting is good, the background is adequate, the cameras are good, they are placed in a way so that we can get the best pictures that we need to get? is there anyone in charge of that? >> cbp would be and that's going to be based on our results of, say, the match rates and you can have an airport and the windows are such that the sunlight comes in effect these boots during the morning in these boots in the afternoon, and those of the things we've got to look at as we deploy this, one of the environmental factors that are going to influence all the different locations that we are going to do this and then we try to adjust, that might mean we have the airport to the windows. >> how do you do that? >> we do that internally by reviewing the data and the results. >> on what kind of basis? weekly? daily? monthly? you know what ten comes in that window, you know what time the sun comes in that window,. >> i would say we do it continuously to get the best production we can out of it. >> thank, you thank you very much. thank, you mister chair. >> thank you very much, the chair recognizes the ranking member. >> thank you, mister chairman, i have a unanimous request, unanimous consent request that two articles on the topic be entered into the record. >> without objection. >> so. ordered >> the chair recognizes gentlelady from texas, miss jackson-lee. >> let me thank the chairman very much and let me acknowledge all of the witnesses and thank them for their service. >> let me renew the inquiry being pursued by chairwoman rice, but i will add to it and that is, ms., wagner better understanding, we will provide the information to the committee on the denial of trusted traveler as it relates to states that may not have the laws that you think are appropriate or, in the instance of new york, closing out access to the issue of drivers licenses, and i raise the question because we should look, as a federal government, at what other intensification options may be valid and i know that we've known each other for a long time, and i think he would be willing to look at that so we can find common ground. on a pursue this line of reasoning and please, witnesses, understand that i am not saying this is what you are doing, i need to understand you are thinking, so, to the deputy executive assistant commissioner wagner, would you accept the fact that bias could be introduced by technology in the application developer of the program has a heavy bias into how it applies to different types of people, because it is technology? >> yes. >> and i would also make the point, that is a little bit humorous, which is a i'm sure the people in iowa were trusting of the app, what they had something going on there and we all can see where we are at this point. would you accept that, mr. mina? and would you accept that, mr. ruin? >> yes. >> and, an algorithm could be made to flag all black males wearing dreadlocks, as wagner? this is in terms of how technology can be. >> i guess you could. >> i understand. and you could say on the record to your knowledge you're not using that kind of -- >> were not using that. >> that will be very sure, i am sure dreadlock whereas would be glad of that. mr. mina? >> yes, that is possible and again, like mr. reiter said, we have not seen that. >> all right, mr. remain? >> it is certainly possible. >> all right, so, here we are, my colleagues over here our dna advocates, as a member of the judiciary committee which had to step away from, we are dna lovers, or the violence against women act and put in 291 million dollars for dna and handsome, and so we understand that the new added technology, but as the department of homeland security, we made a commitment post 9/11, and george bush going to the trade and saying he heard the firefighters been at the same time, he also heard muslims were indicating it is not the blanket world where people who happen to be muslim so, in particular, mr. mina, i want to try to find out what aggressive role you play in helping to not have platitudes, forgive me, i'm not suggesting you do but too aggressively ensure that the biases against black women with dreadlocks, members dreadlocks, muslims, six, wearing a, tire we've been through, that it be on this committee since the beginning, that is not technology but that the now sophisticated technology is not undermining the civil liberties and civil rights of this nation and those coming in innocently to the country, you can use the new technology throughout, and then to mr. remain, let me find out, how are you continuing to do your assessment of these algorithms to ensure that, it looks like you were not able to get the exact one that mr. wagner's team is using. that concerns me. i need you to get every accurate piece of information and i'd like to say that mr. mina, what are you doing to protect travelers and the american people? >> thank you for the question, congresswoman. i think we're doing a lot of different things across the spectrum within the life cycle of this program and policy and again i want to focus, our attention is really on the applications not so much of the algorithm itself but on how it is applied by this program, in this case, cbp, we do that through, on the policy making, side working directly with the component, advising on regulations, and limiting policies as well as offering suggestions as it relates to application, for example, folks wearing religious headwear or folks that have objections to photography based on religious reasons or people who are disabled or other was injured and are not able to take pictures, we also do it through our robust community, again, we talk to members of the community across the country, mister chairman, i actually have the information from me regarding some of the areas, the issue has been raised in portland, atlanta, chicago, seattle, and also to a lesser extent in southern california, and the new york city area stakeholders, our concerns regarding official recognition technology and one of our primary roles is to be the eyes and ears the department and being for our colleagues at cbp, i dhs absentee, here are the concerns we are seeing. how do we work together to try and address some of these problems, or potential problems before they have even greater effect, and then also, we have a robust compliance process, and what we don't have an active investigation right now, that is always something we are looking at is, if we see an issue, we will almost certainly open investigation advice. >> as he, answers mister chairman, and you want to see this on the record, if we can get answers from mr. wagner about what is storms repeating information. >> miss jackson, doctor remain you can answer the question and submit in writing. >> thank you. mr. romine, my question was -- >> a bigger pardon him. >> my question was what are you doing to be accurate in your testing, and you said you did not know whether you had the accurate method they were using. what are you doing to be aggressive making sure that we don't have bias in these? >> yes, ma'am, the test that we undertake are intended to determine whether there are demographic differences, commonly called bias. the fact that i know there is strong interest in testing with data that is more representative, and we saw in a recent mou with the cbp to undertake continued testing to make sure we are doing the very best that we can to provide the information that they need to make sound decisions. >> thank you very much. i yield back. thank you. >> the chair recognizes the gentleman from texas, mr. green. >> thank, you mister chairman. i thank the witnesses for appearing as well and i would like to address some intelligence that has been afforded me. the indication is that and i asked he found that asian african american faces were ten times more likely, ten to 100 times more likely to be misidentified than white faces and i'm curious as to whether or not there is something inherent in the technology that creates an inverse relationship with reference to the identification of whites, juxtaposed to african americans and asians. is there something here in the technology? meaning, if you want to absolutely identify whites, will there be something that you cannot adjust such that you will get the same absolute identification with minorities, asians, african americans, or if you want to, absolutely identify african americans and asians, will you as a result of technology, not be able to properly identify whites? >> a very interesting question, congressman, and let me clarify first that those differentials that we observe are not in the case of identification but rather verification, the one to one testing rather than the one to many testing in general. we saw those demographic differences for african american and for specific islanders and asians as well, that, but in the case that you're talking about, our workers not to date focused on cause and effect, what is it that is causing the algorithms to exhibit certain kinds of behavior? we are really just testing the performance. i don't know the answer to your question. >> my question, was interesting, as you put, it your answer was intriguing, because this is not the first opportunity for the word to be heard that we had these difficulties and at some point it would seem that you would move from testing technology as it is to understanding why technology works the way it doesn't help me to understand why we haven't made that move. >> the question that you asked is a very challenging open research question but we do have some indications, there are algorithms that have been submitted to our testing from asian countries that do not exhibit the demographic differentials on asian faces so we can't guarantee, but we think that's an indication that the training dated that are being used for the algorithm development may have a significant impact on their ability to exhibit demographic differences for different populations. >> do you believe it is important for us to move expeditiously to answer this question, to resolve this issue? such that we don't find ourselves having deployed something en masse that will be notably defective or have some degree of inefficiencies associated with it, the advocacy of this is important. >> yes, sir. i think those are two different things to think about. the performance testing that we currently execute can help operational agencies ensure that they are not deploying things that exhibit demographic to differential -- differentials. the research question that you teed up is fascinating about what are the causes of the demographic differentials. this is a much deeper deeper question a much more difficult, i think. >> well, is it fair to say that the country is, how about 45 seconds left, but the countries that includes technology and indicated to you they are having fewer challenges? is it fair to say that that technology also captures white men sufficiently? >> in the testing that we did for the specific one that i'm referring, to the high performing algorithms from asian countries that don't exhibit the demographic differences on asians, it's in comparison to caucasian faces that i make that statement so there's no difference in the performance, on caucasian faces and asian faces from certain asian developed algorithms and when speculation is maybe the training data that are used. >> thank you, mister chairman. i yield back. >> thank you very much, the chair recognizes the gentleman from rhode island for five minutes. >> thank, you mister chairman, i want to thank our witnesses testimony here today, and thank you for what you're doing, you're dedication to this issue, i certainly believe the technology is important, some of our most vexing issues and challenges, including how to manage an ever growing number, i think it is been a good discussion here today, what i want to ask, if mr. wagner or mr. miner, we know that in technological solutions such as visual recognition software, the outcomes are only as good as the data that inform them so i want to know how his cbp adjusted or augmented the data that it uses to train its software, and what are you doing to ensure the software is being updated as more robust did assistant algorithms are incorporated and training? >> at work closely with the vendor whose algorithm we are using, and we work more closely with them to incorporate their updates and their latest and greatest products into how we are using them and then as we review the data we look to make this operational just months which do impact metrics and again that's gonna be the quality of the photograph, the quality of the camera, the human factors, the size of the gallery is really important, and on the margins of the capabilities of these algorithms, it's a couple of thousand and interesting correlation's are how much better improved is your metrics in what is the impact on any potential demographic biases on a much smaller -- smaller gallery or sample size. what we were getting earlier with these variables that we can raise and lower to adjust some of what the air rates are showing us, so. >> so that point then, how to cbp incorporate feedback from officers about arrows that facial recognition software made in the field? because the machine, it learns, and when the officer interacting with someone and the software does not get it correct, unless that feedback they spied back into the system, this is dumb does not learn. >> absolutely and that's where we look at, you, know we look at system logs themselves, we also talked the officers in a provide the feedback and then were also on site to witness and observe and discuss with those officers, as we deploy this. it's important. >> i understand the trust program -- to travel approaches information with other countries. how does it shared biometrics formation with other countries and what steps does it take to ensure that those countries use the data responsibly? is that accurate, number one? my understanding is, and how are we guarding that data to make sure they're protected? >> i'm trying to think of when -- i'm not aware of how we would share or if or even sharing. >> the trusted traveler program. >> we don't share -- we might stick share persons status their proved in a program, but were not actually sharing, and say, their fingerprints. >> so let me ask that one for the record, and i ask that you get back to me, and that this is important. what types of information do we share under the trusted traveler program and i think that's important for us to know, whatever information we share, i want to know what information steps we take to ensure we use that data responsibly. i'm going to go -- iowa losses in a different way, and just prior to our hearing on this topic, we notified of a cyber incident of the network of a cbp subcontractors, someone complaining -- claiming to be a foreign agent gain access to tens of thousands of photos, of drivers faces a license place at a port of entry along the southern border. how does cbp -- that was cbp ensuring personal did it collects for facial recognition technology screen programs whether by the government direct to do all by vendors or other private sector partners are being protected by inadvertent or otherwise unauthorized access and also, what insurances can you give our committee that the root causes of the may 2019 have been addressed so as to reduce another breach. but the airlines we have designed a set of we have assigned set of efficiency requirements which they can get to, not storing, not sharing, not saving any of the photographs when i take the picture, they have to transmit to us and purged from their system. one of the other conditions is that we have to be available for cbp to audit cameras in technology to ensure that they are following those rules. we are about to commence an audit on one of the airlines in the next couple of months and start that process to do that but to make sure that that's not happening. >> thank you. mister chairman i would ask that miss wagner get back to the committee in writing as soon as possible about the trusted traveler program and what information is shared with partners. >> okay. >> thank you mister chairman. >> thank you. the gentlelady from texas wants to ask a question. >> thank you so very much, first of all, i would ask unanimous consent to plays into the record, not the witnesses but the headline reads, amazon facial recognition mistakenly confused 28 congressman with known criminals, july 26th, 2018. mr. wagner, does want to use, are using the amazon technology? >> we are not using their matching algorithm. >> thank you, mister chairman, my question, is you gave mr., a congressman a detailed response, so let me try to change it around to, do you have a team that is directly responsible not just for the implementation but for the internal analysis of the utilization of the technology that you're using so that it is on site, so that you're able to be firsthand knowledge of the violations or whether these were abuses by way of the technology, is that information coming back to your office, when you see that your sector. >> yes, part of it is our office who does that working in conjunction with few locations. >> so do you have a team that is just responding to that? >> we have teams that review the data, the functioning of the systems, review the compliance of the officers using the technology, yes. >> i know we have a lot of, work you do a lot, maybe they can be a classified briefing, i would like to do a deeper dive on how that is done and how that is kept and whether they store they keep the data on mr. jones and mr. various named persons. >> we will work through it. >> the data is all stored in compliance with the systems of record notice is where that data is stored so the photograph of the u.s. citizens we take in 24 hours has purged. a picture of a foreign send over to the department database where destroyed four 75, years the record at the border crossing, the biographical information is stored in other systems,. we >> will follow up on your quest. >> thank, you witnesses. >> let me insert in the record a letter from the electronic privacy defamation incest are and a pressure to the u.s. travel association and mr. wagner, if you are aware of any notification requirements that a state would be noticed, and i'm talking about global entry situation, it looks like new york was the first of one or two others, since we've been sitting here, mr. cuccinelli has said washington state might be in a similar position, and i'm just wanted to make sure that if this is a way forward then surely, in light of what what mr. rosen some other new yorkers on this committee have said, there should be some notice that this is about to happen, not just a press conference. i think if you are aware of any, please get it back to us in the committee. we love to have it. i think -- i thank the witnesses for the valuable testimony and the members for their questions. the members of the committee may have additional questions for the witnesses and we ask that you respond expeditiously in writing to those questions. without objections, a committee record will be kept open for ten days. hearing no further business, the committee stands adjourned. [inaudible conversations]

Related Keywords

New York ,United States ,Louisiana ,North Carolina ,Texas ,Iran ,Lebanon ,China ,Rhode Island ,Illinois ,California ,Washington ,Canada ,Michigan ,Seattle ,Germany ,New Jersey ,Iowa ,Chicago ,Americans ,America ,Iranian ,German ,American ,Alaskan Indian ,Mister Higgins ,Mister Jean Wagner ,Las Vegas ,Mister Wagner ,Chuck Roman ,