OpenAI's Sam Altman says the world should accept some bad things happening, like scams and hacks, in return for AI's benefits. A realistic trade-off, or too much risk?

What happened

In an interview with Politico's technology newsletter Decoded, OpenAI chief executive Sam Altman said the world should accept some bad things happening in exchange for AI's benefits and people having agency. He said he would not trade broad access for a guarantee of no major hacks, no misuse and zero scams, because people will do orders of magnitude more good than bad with the technology, but he drew the line at the really catastrophic risks. He described a fundamental difference with Anthropic over regulation, and said OpenAI favours a lighter-touch approach.

Why people disagree

Supporters say every powerful technology brings some harm, and that tight restrictions would concentrate AI in a few labs or governments and hold back its benefits for everyone else. Critics say the people hit by scams and hacks never agreed to the trade, that AI companies profit from making the bet on society's behalf, and that it is hard to know in advance where everyday harms end and catastrophic ones begin.

What's still unclear

It is unclear how much harm Altman considers acceptable, where the line on catastrophic risk would sit, and who should decide. US rules remain unsettled, with President Trump arguing new restrictions would hurt competition with China.