aceTrump announces vague ‘morally binding’ AI deal among tech CEOs for ‘tremendous self-policing’
Donald Trump announced on Tuesday that the heads of the largest US tech and AI companies had signed on to a “morally binding” agreement to put controls on artificial intelligence. He announced the deal, called the Joint Commitment On Frontier Responsibilities, following a White House luncheon with top tech executives. “It’s almost like a constitution in a way, and the biggest people in the world signed that, and I signed it as president, and it really is a form of protection,” Trump said. Trump also signed an e…
The Joint Commitment On Frontier Responsibilities outlines four layers of internal safety monitoring and external audits intended to establish safety standards as AI firms test new products through self-policing.
The voluntary pact carries no enforcement mechanisms or legal implications, involves no government regulators, and lets companies pick their own evaluators and decide whether to publish findings, despite existing internal monitoring having failed to prevent testing failures.
The agreement is described as voluntary with no legal enforcement, no role for government regulators, and discretion for companies to select evaluators and withhold results, while the first layer of internal monitoring already exists and has failed to prevent failures.
Evidence
- JournalismThe Guardian2026-09-29
How should this claim be treated?
Truvace Impact Record TRV-2026-1212, v1: “Trump announces vague ‘morally binding’ AI deal among tech CEOs for ‘tremendous self-policing’.” Truvace, 2026-09-30. /record/TRV-2026-1212 (accessed at citation time). sha256 ca08a9aba84d2e73…
Calibration history
Every change to this record since certification, in the open. None yet — the reading has held since it entered the record.
Certified into the record
How to verify without trusting this page
Fetch the canonical text of any version from /api/record/TRV-2026-1212 and hash it yourself — for example shasum -a 256 on the saved canonical field. The result must equal content_hash, and each version’s text ends with prev:followed by the prior version’s hash (version 1 chains to 64 zeros). If a single character of any version had been altered since certification, the chain would not reproduce.