AIBRIEF
Dario Amodei dined with Trump and the leaders of Google, Meta, OpenAI, xAI and NVIDIA to sign a 'morally binding' AI safety accord. The published text does not describe an enforcement mechanism.
Anthropic CEO Dario Amodei joined a White House AI safety dinner on September 29, 2026, alongside Trump and the heads of Google, Meta, OpenAI, xAI and NVIDIA. Trump publicly called out "Dario has been fantastic" during the event. The published seating chart placed Amodei three seats away from the president.
The accord Trump and the six companies signed that night is what the White House billed as a "morally binding" AI safety framework. As reproduced in commentary by Zvi Mowshowitz, it commits participants to internal controls, an internal oversight team, independent external auditors, and an independent oversight committee. Christian Robles reported the same four-layer arrangement for Nextgov, drawn from a copy of the accord.
The accord text does not establish an enforcement mechanism or require external audit publication. No demonstrated safety improvement is shown. A separate September 29 executive order directs federal terminology changes and a proposed legislative definition within 60 days. It does not establish the four-layer structure.
The dinner bought Anthropic and the other signatories permission to keep engaging a hostile administration on AI safety terms. It did not buy a safety milestone.