Breaking

Trump and Six AI Leaders Sign Voluntary Frontier-AI Safety Accord

October 1, 2026 at 18:23 EDT

Policy & Regulation
Foundation Models
Security

President Donald Trump and leaders from six major AI companies signed a voluntary safety accord on September 29, putting industry self-regulation at the center of the U.S. response to frontier-model risks. The pact commits the companies to internal controls, outside evaluation and board oversight, but it does not create binding federal rules or penalties.

Sep 29, 2026 | U.S. AI policy

Voluntary AI safety accord

Six leading AI companies agreed to a self-regulatory pact built around four layers of controls, audits and board oversight.

Signatories

Control layers

Definition proposal

60 days

The pact by the numbers

6 companies signed

4 oversight layers

Four-layer safety loop

1. Internal controls

Monitor capabilities and alignment.

→

2. Internal team

Check controls and fix issues.

→

3. External audit

Test whether safeguards work.

→

4. Board oversight

Review reports and remediation.

Supportive read

A shared checklist could normalize monitoring, external evaluation and board accountability without pausing development.

Critical read

The pact is voluntary and does not itself create penalties, public audit findings or binding federal duties.

What remains unsettled

Congress, regulators and the administration still have to decide whether these self-policing commitments become enforceable rules.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.