Urgent.News

What's breaking now, across thousands of outlets.

AI

Trump's tech advisor has advice for AI companies worried they're building unsafe models: 'Just stop'

AI companies have called for more regulation of their own products to ensure safety. Trump's tech advisor says they can slow down whenever the want.

Michael Kratsios, the director of the White House Office of Science and Technology Policy, has advised AI companies to simply halt their development if they believe their models pose a risk. In an interview with Fox News Sunday, Kratsios stated that AI companies do not require government permission to do so. The Trump administration has been supportive of expanding and accelerating AI infrastructure.

Kratsios dismissed the notion that the US government needs to intervene for companies to take a pause, emphasizing that companies can simply choose to stop if they deem their technology unsafe. He pushed back against the idea that the government must take explicit action, stating that the government possesses various tools to address such issues as they arise.

The concerns surrounding AI's rapid development center on potential public safety risks, including job loss, privacy, and cybersecurity threats. OpenAI CEO Dario Amodei recently expressed worry about the industry's acceleration, suggesting that independent safety monitors should be placed within AI labs and calling for more robust government regulation.

Written by urgent.news from Business Insider's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at businessinsider.com →

More in AI

Researchers penetrate OpenAI systems with Claude models

Security researchers participating in OpenAI’s bug bounty programme penetrated employee accounts and reached the company’s private GitHub monorepo after using Anthropic’s Claude models to develop an…

  • Researchers from Hacktron AI exploited OpenAI systems using Claude models.
  • Memory-safety flaw in libheif library enabled remote-code execution.
  • Claude Opus 5 used to create proof-of-concept pull request in OpenAI's repo.

More from Sunday 20 September →