Trump and Six AI Chiefs Sign a 'Morally Binding' Voluntary AI Accord

President Donald Trump said he and the heads of Anthropic, OpenAI, Google, Meta, xAI and Nvidia had signed a voluntary accord committing their companies to 'robust internal controls' on AI. Trump called it 'almost like a constitution' and repeated that he would not 'stifle' the industry; critics called the pledge vague and unenforceable.

Key Takeaways

President Donald Trump said he and the heads of Anthropic, OpenAI, Google, Meta, xAI and Nvidia had signed a voluntary accord committing their companies to 'robust internal controls' on AI.
Trump called it 'almost like a constitution' and repeated that he would not 'stifle' the industry; critics called the pledge vague and unenforceable.

AI synthesis4 sources6

President Donald Trump said on Tuesday that he and the leaders of the largest US artificial intelligence companies had signed a voluntary accord on how their models are developed, describing it as a 'morally binding' commitment built on 'robust internal controls' rather than government regulation.

Trump unveiled the agreement after a White House meeting with the industry's biggest names. According to Al Jazeera, the signatories were Dario Amodei of Anthropic, Greg Brockman of OpenAI, Sundar Pichai of Google, Mark Zuckerberg of Meta, Elon Musk of xAI and Jensen Huang of Nvidia.

According to accounts of the document, the agreement calls for both internal and external reviews of the companies' AI work. Trump framed the pact as a statement of shared principles and said it was 'almost like a constitution', while repeating his position that his administration would not 'stifle' the sector. He has previously rejected binding, industry-wide guardrails, favouring a light-touch approach intended to keep the United States ahead of China. Politico reported that he also floated a ten-person committee that would play a role in overseeing the technology.

Critics were quick to note what the accord does not contain: no independent audit, no penalties and no statutory force. Because it is voluntary, companies remain free to set their own testing and disclosure standards — the gap safety researchers have warned about repeatedly as models grow more capable.

The announcement comes days after OpenAI shelved its GPT-6.1 Astra model over internal safety findings, and amid continued debate over how far frontier AI development should be allowed to run ahead of oversight.