Advertisement

Add GIZMOCHINA as Preferred Source on Google

Sam Altman isn’t pretending that AI will be perfectly safe forever. Speaking at Salesforce’s Dreamforce conference, the OpenAI CEO said he believes a certain number of accidents with emerging technology are simply unavoidable. What matters most, he argued, is how the industry responds when they happen.

“I do think that some degree of accidents are unavoidable with new technology,” Altman said. “What I really care about is that we have got a great culture of accident reporting and learning.” He pointed to the Federal Aviation Administration and the National Transportation Safety Board as models. Those agencies dig deep into every incident, and that culture of learning is a big reason modern air travel is so safe.

His comments come at a time when people are watching leading AI labs more closely than ever, especially as systems get closer to greater autonomy. Recent remarks from Jacob Coxon, a researcher who recently left Anthropic after previously working at OpenAI, added fuel to the debate by warning that companies are racing toward self-improving superintelligence.

Altman didn’t dismiss the worries. He said public concerns about AI getting out of control or power becoming too concentrated in a few labs are completely understandable. “The world is right to be afraid of this,” he said.

However, instead of slowing everything down or stopping, Altman argued that companies need to keep work on alignment, safety, and monitoring well ahead of the advances in model capabilities. If they can’t maintain that balance, he said, they should be ready to slow or even pause research. The point isn’t to shrug and accept disasters. It’s to be honest when things go wrong, review them quickly, and keep improving so the risks don’t spiral.

These comments fit into a larger ongoing conversation about how fast frontier AI should move. Some industry voices have been calling for more caution, while US policy under the current administration has focused on staying ahead technologically.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

Comments