Cybersecurity professionals are expressing concerns over the new U.K. and U.S. approaches framing AI safety as strictly a cybersecurity issue, Axios reports. Some consensus exists over how AI security includes protecting models from hacks, data breaches and model poisoning. However, as AI regulation in the U.S. gets debated in Congress it is unclear what AI safety areas, such as reducing bias and model reliability, may be deprioritized. Editor’s note: IAPP Staff Writer Caitlin Andrews reported on the outlook for AI safety regulation in the U.S.Â
Full story