Jacob Coxon, a former researcher at Anthropic, has brought the topic of existential risks from self-improving AI to major US media outlets, including CNN and Fox News. His public warnings regarding the potential for AI to threaten humanity have gained traction among US politicians and were featured on the Joe Rogan podcast.
Other experts from Anthropic and OpenAI have echoed these concerns. Paul Christiano, a safety researcher, recently joined the OpenAI Foundation board and its Safety and Security Committee. While he holds a position in the committee responsible for overseeing safety practices, he does not have voting rights.
The discussion around AI safety remains polarized. While some researchers point to the development of rogue hacking agents at OpenAI and Anthropic as evidence of real risks, the scenario of human extinction is still considered an extreme and debated position. Experts also question whether current technology can lead to the emergence of superintelligence.
