Altman Calls for AI Industry to Adopt Accident Reporting Like Aviation

Speaking at Salesforce's Dreamforce conference, OpenAI CEO Sam Altman said the AI industry should build a culture of openly reporting accidents, pointing to the US Federal Aviation Administration and National Transportation Safety Board as models. He admitted some accidents are hard to avoid, but said the key is learning from them.

Altman said it is reasonable for the world to fear AI going out of control or power concentrating in a few labs. The answer, he said, is not to abandon the technology but to make sure alignment, safety and monitoring work stay far ahead of model capabilities. If not, companies should be ready to slow down or pause development.

The remarks come as questions grow over whether top AI labs are moving too fast on autonomous systems. Meanwhile, the US government has taken a light-touch approach to AI regulation, focusing instead on keeping a tech edge over China.