Trump Administration Brings Tech Giants Together on Non-Binding AI Safety Accord
Developing story first seen 1 hour ago
The six signatories—Meta, Google, OpenAI, Anthropic, Nvidia and xAI—have agreed to a non-binding White House accord on AI safety, published by Donald Trump on Truth Social. It calls for companies to monitor their models, check their safeguards internally and through external evaluators, and have board-level oversight, but leaves key standards and inspection schedules undefined.
The accord says the companies will meet regularly to develop safety standards and best practices, while leaving open the possibility of future laws or regulations. Separately, Trump signed an order directing federal agencies to use “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI”; his science and technology adviser has 60 days to work with agencies on a definition.
- Six AI companies signed a voluntary White House safety accord.
- The accord sets out audits and oversight but leaves standards vague.
- Trump ordered federal agencies to adopt “Super Intelligence” terminology.
New here? Start with this
Artificial intelligence systems have become increasingly powerful, but the rapid pace of their development has raised concerns about safety risks and potential misuse. Policymakers and technology companies are seeking to manage these risks while allowing AI innovation to proceed.
The Trump administration has brought together six major technology companies—Meta, Google, OpenAI, Anthropic, Nvidia and xAI—to sign a voluntary accord on AI safety. The accord commits these companies to monitoring their AI systems for problems, testing their safety through both internal reviews and independent external assessments, and ensuring their boards have oversight of these efforts.
The accord is non-binding, meaning companies are not legally required to follow it and several key details remain to be worked out, such as what specific safety standards will be required and how often companies will be inspected. The companies have agreed to meet regularly to develop shared safety practices, and the accord allows for the possibility that new laws or regulations might be introduced in the future.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Supporters of this accord contend that industry-led self-regulation harnesses genuine technical expertise and market incentives where government bureaucracies cannot. The signatories have substantial reputational and legal exposure should their systems fail, making safety alignment economically rational. Board-level oversight signals serious commitment, and the accord's provision for regular collaborative development of standards allows practices to evolve with the rapidly changing technology rather than being locked into potentially outdated regulations. This framework preserves space for innovation whilst establishing baseline safety expectations.
The case against
Critics argue that non-binding agreements are inherently unenforceable and amount to corporate public relations rather than substantive governance. Tech companies' historical record on self-regulation—from privacy to content moderation to algorithmic harms—demonstrates consistent prioritisation of commercial interests over safety when the two conflict. The accord's vague language, undefined standards, and absence of inspection schedules or enforcement mechanisms mean compliance cannot be verified or compelled. Placing regulatory power in private hands circumvents democratic accountability and renders safety oversight dependent on corporate goodwill rather than law.
Coverage
- BBC Technology — Trump gathers tech executives to discuss AI safety and expansion
- The Register — Trump administration gets Big Tech to sign weak, non-binding, AI regulations