Altman defends broad AI access despite accepting some risks
In a new POLITICO interview, OpenAI chief Sam Altman argues that AI’s benefits and people’s agency justify accepting some harmful outcomes while keeping the technology broadly accessible. He describes a continuing difference in worldview with Anthropic, even as the companies converge on specific safeguards. Public access remains central to the trade-off he advocates between benefits and risks.
Artificial Intelligence··Morning
Altman links accepted risks to public access and benefits
Sam Altman, chief executive of AI company OpenAI, argues in a new interview that society should accept some harmful outcomes in exchange for AI’s benefits and people retaining agency. He also says the technology should remain broadly accessible. POLITICO’s Decoded interview, published on October 4, and SiliconANGLE’s account both describe that position. These are Altman’s views about the trade-off, rather than evidence that a particular system is safe.[1], [2]
Worldview differences remain despite agreement on safeguards
Asked about rival AI company Anthropic and its chief executive Dario Amodei, Altman says a substantial difference in worldview remains. POLITICO sets that answer against the companies’ closer policy positions. Anthropic has long advocated AI regulation, while OpenAI previously preferred a lighter approach. More recently, OpenAI has backed stronger requirements amid reported hacking incidents and unwanted model behaviour.[1]
OpenAI backs stronger requirements in specific proposals
Altman agreed with Amodei’s call the previous month to slow development of the most advanced models. OpenAI has also endorsed state laws with stricter safety requirements than its earlier strategy supported. Its lobbyists recently backed a bipartisan House proposal requiring leading AI companies to include external safety evaluators.[1]
