00:00It's Benzinga bringing Wall Street to Main Street.
00:02Google DeepMind CEO Demis Hassabis proposed an international watchdog to test and review
00:09advanced artificial intelligence models before release, according to Bloomberg. The group would
00:14include independent technical experts, receive funding from leading AI companies, and review
00:19models for up to 30 days. Hassabis said concerns about the cybersecurity capabilities of Anthropics'
00:25mythos model prompted him to publish the plan. He also warned that future models could create
00:31biosecurity risks. OpenAI CEO Sam Altman and Elon Musk praised the proposal, while Hassabis plans
00:37to discuss it with U.S. policymakers in Washington. The watchdog could coordinate an industry slowdown
00:43if risks increase. Hassabis acknowledged political and regulatory divisions could prevent the plan
00:48from advancing, but said momentum may be building for international oversight. For all things money,
00:54visit Benzinga.com.
00:55Benzinga.com
Comments