Alphabet's Google (GOOG, GOOGL), Anthropic, Meta (META), Nvidia (NVDA), OpenAI and SpaceX's xAI (SPCX) agreed to implement a four-layer control and audit system for artificial intelligence under an accord signed Tuesday with US President Donald Trump.

The White House Accord on Super Intelligence calls on the companies to establish internal controls to monitor capabilities and alignment of their AI models during training and deployment and to ensure they are not hacking or accessing technical systems in unintended ways.

Signatories also agreed that each company should establish an internal team to ensure its controls, monitoring and detection systems operate as intended.

Each company should also work with independent external auditors or evaluators to carry out assessments of these systems, according to the accord.

The fourth layer calls for each company to establish an independent board committee to oversee and receive reports from internal teams operating these controls as well as from auditors and evaluators.

The companies also agreed to meet regularly to establish standards and best practices aimed at improving the safety of their systems.