Trump and Major AI Firms Sign Voluntary Pact on AI Safety Controls

Trump and leaders of OpenAI, Google, Meta, xAI, Nvidia and Anthropic sign a voluntary pact on AI safety controls and oversight

President Donald Trump and leaders of major artificial intelligence companies have signed a voluntary agreement establishing safeguards for advanced AI systems, including measures aimed at cybersecurity, biological and chemical threats, and unauthorized access to computer systems.

The agreement involves major technology companies including xAI, Meta, Nvidia, Google, Anthropic and OpenAI. According to the details cited by Coin Bureau, participating companies committed to establishing stronger internal controls, allowing independent outside auditors to assess those controls and creating internal oversight mechanisms.

The accord is not legally binding. Trump described the commitments as “morally binding,” according to the post.

Companies Commit to Four Layers of Oversight

The agreement calls on participating AI companies to put controls around their models to monitor risks involving cyber, biological and chemical threats. It also addresses the possibility of AI systems accessing or interfering with computer systems without authorization.

Companies are expected to establish internal teams responsible for monitoring these controls, alongside board-level committees that oversee the broader safety process. Independent external auditors are also expected to assess whether the safeguards are operating as intended.

The framework places substantial responsibility for implementation within the companies themselves. The voluntary structure means the commitments do not carry the same legal enforcement mechanism as mandatory federal regulation.

Voluntary Framework Sets Out Self-Governance

The agreement comes as U.S. policymakers and technology companies continue to address the risks associated with increasingly capable AI systems.

The White House has previously pursued voluntary cooperation with AI developers. In June, an executive order directed the development of a voluntary framework through which AI developers could work with the federal government on cybersecurity involving covered frontier models. The order explicitly stated that the framework would markets not create mandatory government licensing or pre-clearance requirements for AI models.

Individual companies have also developed their own safety systems. Anthropic, for example, maintains a Responsible Scaling Policy covering risks associated with increasingly powerful models and has documented safeguards addressing cybersecurity as well as chemical and biological threats.

Oversight Remains Voluntary

The latest agreement therefore establishes a common set of commitments among participating companies without making them legally enforceable requirements. Its effectiveness will depend on how the companies implement the controls, how independent audits are conducted and how oversight committees operate.

For now, the central commitments are the creation of internal safeguards, independent external assessments and board-level oversight covering risks from advanced AI systems.

Trump has characterized the agreement as “morally binding,” while its formal status remains voluntary rather than legally mandatory. The participating companies are expected to carry out the safeguards through their own internal governance structures.

Writer: Victoria Hale
Technology & Blockchain Writer

Victoria Hale writes about blockchain technology, digital infrastructure, and the intersection of emerging technologies with finance. Her articles explore how new protocols and systems are shaping the evolving digital economy.

She prioritises clarity and accuracy when explaining technical developments to a general audience.