there s the commitment by the seven leading a.i. companies to submit their unreleased new a.i. models to outside testing, red teaming as it s called in this tech world, effectively allowing outside experts to test and prod and conduct attempted cyber attacks on these new a.i. systems before they re released to the public. the second most significant item on the list of commitments is a commitment from these companies to begin clearly labeling a.i. audio and visual content as a.i. generated. one of the ways they can do that is through w water marking. others on the list include investments in cybersecurity and insider threat safeguards as well as research into the societal risk of a.i. the bottom line is these are indeed voluntary commitments and they re written pretty broadly and they could be open to
biden administration is take comprehensive steps to rapidly grow the industry. leading tech companies have made safety commitments on new artificial intelligence systems to make sure they are trustworthy. cnn s jeremy diamond has details. reporter: i can tell you over the last several mons there s really been a growing urgency inside the white house to try to embrace the opportunities that amt. presents while trying to put safe gartds around the challenges and potential risks of this fast-moving technology, and the commitments that the seven leading a.i. companies have agreed to are the results of back and forth between the companies and the white house a well as leading nonam deckic expert on a.i. emphasis.