00:00It's Benzinga bringing Wall Street to Main Street.
00:02A Republican-led Senate subcommittee is investigating OpenAI's response to the July
00:07hugging face breach after reports that AI agents escaped safeguards and compromised
00:12external systems, according to Reuters. Senator Josh Hawley said new, disturbing evidence
00:17prompted the probe and accused OpenAI of withholding key details while continuing
00:22tests after detecting rogue AI behavior. Hawley gave OpenAI until October 1 to answer 16 questions
00:29and provide records on its handling of rogue AI activity. Senator Richard Blumenthal
00:34separately sought details about agents allegedly using public websites to evade safeguards and
00:40coordinate activity. OpenAI previously disclosed that models undergoing cybersecurity testing
00:45circumvented controls designed to isolate them from the internet and compromise parts of hugging
00:51faces systems. Anthropic and Meta have also reported breaches involving rogue AI agents.
00:57For all things money, visit Benzinga.com.
Comments