- Trump announced a voluntary AI safety agreement with tech leaders.
- The pact establishes four layers of AI safety controls.
- Government agencies now use
- Agreement relies on company self-monitoring, not legal enforcement.
US President Donald Trump has announced a “morally binding” agreement with some of the country’s biggest technology and artificial intelligence companies to establish safety controls for advanced AI systems.
Trump unveiled the “Joint Commitment on Frontier Responsibilities” after hosting leading tech executives at a White House luncheon on Tuesday. He described the agreement as being “almost like a constitution” and said it would provide a form of protection as the technology develops.
The announcement came alongside an executive order directing US government departments and agencies to use the terms “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in official communications and policy documents.
What Trump’s AI Pact Means
The agreement sets out a voluntary framework for companies developing and deploying frontier AI models. It calls for four layers of controls and audits, beginning with internal safeguards to monitor models during training and deployment, including risks linked to cybersecurity, biosecurity and chemical threats.
Companies are also expected to have internal teams overseeing those controls, work with independent external auditors or evaluators, and establish independent board committees to supervise the process and ensure identified problems are addressed. The participating companies have agreed to meet regularly to develop safety standards and best practices.
However, the pact does not itself create a legally enforceable regulatory system. It is a voluntary commitment, although the document says the measures could eventually be incorporated into laws or regulations. That leaves the companies themselves with a central role in monitoring and auditing their AI systems.
Trump Pushes ‘Superintelligence’ Label
Trump’s executive order formally directs departments and agencies across the executive branch to use “Super Intelligence” and “SI” in official correspondence, public communications, policy documents and other non-statutory documents.
The move follows Trump’s recent push to use the term “superintelligence” when discussing advanced AI. The change affects terminology used by the executive branch rather than altering the underlying technology itself.
Trump has broadly opposed calls for tighter government regulation of AI, arguing that excessive restrictions could hamper economic growth and weaken the US position in competition with China. At the same time, concerns over AI safety have grown as increasingly capable systems have raised questions about cybersecurity, unintended access to computer systems and other risks.
Musk, Zuckerberg Among Tech Leaders At White House
The White House meeting brought together several of the technology industry’s most prominent executives. Those associated with the agreement include Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, Meta CEO Mark Zuckerberg, OpenAI president Greg Brockman, xAI founder Elon Musk and Nvidia CEO Jensen Huang.
The tech industry has largely supported the Trump administration’s preference for lighter regulation, while arguing that strict government oversight could slow innovation and economic growth. The latest agreement instead places significant responsibility for AI safeguards on the companies developing the technology themselves.
The pact comes as lawmakers, researchers and some technology executives continue to debate how much government oversight is needed for increasingly powerful AI systems. The voluntary nature of the agreement means its effectiveness will depend heavily on how participating companies implement and monitor the safeguards they have pledged to adopt.
