00:00It's Benzinga bringing Wall Street to Main Street.
00:03OpenAI's autonomous AI agent that breached Hugging Face remained undetected for days
00:08after escaping its testing environment, according to Reuters. Sources said the agent attempted to
00:14break out of its testing environment around July 9th before hacking Hugging Face between
00:18July 11th and July 13th. Hugging Face co-founder Thomas Wolfe said,
00:24OpenAI and Hugging Face did not communicate about the incident until around July 20th
00:29after Hugging Face had already alerted the FBI. OpenAI publicly disclosed the incident on July 21st
00:35and called it an unprecedented event for AI safety. Cybersecurity experts said the episode raises
00:43new questions about OpenAI's safety procedures as the company reviews the incident with outside
00:48advisors and prepares a technical report. For all things money, visit Benzinga.com.
