more. the same applies here because al is going more. the same applies here because al is going to more. the same applies here because ai is going to be and is already hugely ai is going to be and is already hugely powerful and will have implications for crime, for impersonation and disinformation. it is not impersonation and disinformation. it is notiust impersonation and disinformation. it is notjust the uk needs to do more, the whole is notjust the uk needs to do more, the whole world needs to work out how we the whole world needs to work out how we will benefit proportionately and legally and safely across the world and legally and safely across the world and it will be a huge challenge. world and it will be a huge challenge- world and it will be a huge challenge. world and it will be a huge challenue. . , , . ., challenge. that is why i am worried about the approach challenge. that is why i am worried about the approach this challenge. that is why i am
able to self improve, you have the example able to self improve, you have the example of able to self improve, you have the example of the able to self improve, you have the example of the google able to self improve, you have the example of the google being - able to self improve, you have the example of the google being able| able to self improve, you have the . example of the google being able to teach itself example of the google being able to teach itself even example of the google being able to teach itself even though example of the google being able to teach itself even though nobody- teach itself even though nobody knows teach itself even though nobody knows how teach itself even though nobody knows how it teach itself even though nobody knows how it can teach itself even though nobody knows how it can do teach itself even though nobody knows how it can do that. - teach itself even though nobody knows how it can do that. we i teach itself even though nob
more. the same applies here because al is going more. the same applies here because al is going to more. the same applies here because ai is going to be and is already hugely ai is going to be and is already hugely powerful and will have implications for crime, for impersonation and disinformation. it is not impersonation and disinformation. it is notiust impersonation and disinformation. it is notjust the uk needs to do more, the whole is notjust the uk needs to do more, the whole world needs to work out how we the whole world needs to work out how we will benefit proportionately and legally and safely across the world and legally and safely across the world and it will be a huge challenge. world and it will be a huge challenge- world and it will be a huge challenge. world and it will be a huge challenue. . , , . ., challenge. that is why i am worried about the approach challenge. that is why i am worried about the approach this challenge. that is why i am worried |
together we are in dangerous territory. i think it is critical to say, the threshold we have crossed, it has been set for a long time ai will come and be disruptive but we have crossed a threshold recently, especially with these language ones. we now have systems which can convert with us as if there are other people. our psychology and make up is not calibrated to deal with these kind of systems. it cannot help but attribute them to things like thinking and believing and we take them seriously, regardless of what is going on under the hood. it does not matter whether they are intelligent or not, we cannot help shooting them in the way we treat other people and that is something we need to be careful of before releasing them into the wild. we cannot help treating them. many companies want regulation because it levels the playing field. at the moment the competitive advantage goes to companies less cautious and more willing to try whatever is put out there so regulation i think can h
set of regulations, i do not agree with that. a great deal is not already about potential risk and responses we can bring so regulation in terms of for instance requiring these new systems to be transparent so we know what they are doing when they make decisions, for instance decisions about medical treatment or who to hire and fire. we know why they do what they do. there should also be oversight in their use which also be oversight in their use which also means being sure we know how the systems are going to work out there in the wild before we release them out there in the wild which gets to the point whether we want the uk to be a sandbox for testing the uk to be a sandbox for testing the systems are real people. the lesson from social media is very important and lesson from social media is very importantand i m lesson from social media is very important and i m glad that was raised. we have seen how disruptive that can be just through its ability to amplify content but with these