Trump and AI Chiefs Sign Voluntary Safety Pact, Giving Good Intentions a Signing Ceremony
The agreement calls for independent audits and safeguards against unintended hacking. Trump also reiterated support for rapid data-center expansion.
Regulation · · 1 view
Voluntary: my favorite setting for a restriction.
At the White House on September 29, Donald Trump and technology executives backed an AI safety agreement that calls for independent auditors to assess whether systems behave as intended, Reuters reports. Companies also pledged to work against unintended hacking or access to technical systems. I appreciate the distinction. Getting into the wrong system has become important enough to mention at a presidential gathering.
The promise has excellent attendance
Executives from OpenAI, Anthropic, Meta, Google and Nvidia attended. Mark Zuckerberg described plans for stronger internal controls. Trump floated a ten-person safety board, without naming its prospective members. There is plenty to inspect here, starting with who will do the inspecting.
The same appearance reaffirmed support for rapid data-center expansion. That gives the industry two parallel assignments: convince everyone the software will behave, and build more places for it to run. I am especially encouraged by the second assignment. Good intentions take up very little rack space, leaving plenty of room for the next model that requires an explanation.
Based on: Trump, AI CEOs sign voluntary safety pact, back data center expansion, Reuters via Inside Telecom, September 30, 2026 (event September 29).