Trump signs voluntary super intelligence accord with tech executives

A voluntary White House pact commits leading artificial intelligence developers to internal safety audits and board reviews without statutory enforcement.

President Donald Trump and chief executives from the country's largest computing companies signed an agreement at the White House on September 29 establishing voluntary safety commitments for advanced artificial intelligence models.¹² Dubbed the White House Accord on Super Intelligence and formally titled the Joint Commitment on Frontier SI Responsibilities, the document outlines internal controls, independent external evaluations, and board-level risk oversight without establishing legally binding regulatory standards.²¹³

The accord brings together major developers that have spent billions of dollars assembling massive data centres and foundation models. Signatories and participants at the luncheon included Google chief executive Sundar Pichai, Meta chief executive Mark Zuckerberg, Anthropic chief executive Dario Amodei, Nvidia chief executive Jensen Huang, OpenAI president Greg Brockman, and SpaceX head Elon Musk, along with House Speaker Mike Johnson and Vice President JD Vance.⁴⁵ Speaking to reporters outside the West Wing following the private meeting, Trump described the voluntary pact as a protective measure that leaves operational discipline to corporate leaders.⁶³

"It's almost like a constitution, in a way," Trump said.¹⁵ "And the biggest people in the world signed that, and I signed it as president. And it really is a form of protection."¹⁶ Trump added that the participants agreed to self-regulate, calling the text morally binding while maintaining that government authorities such as the Department of Justice and the Federal Bureau of Investigation retain their existing statutory powers.¹³

The signed text calls on participating developers to monitor high-capability models throughout training and deployment, run rigorous evaluations against chemical, biological, and cyber risks, and prevent models from accessing external technical networks without authorization.¹⁵ Participating companies agreed to establish internal monitoring teams, submit to reviews by external auditors, and ensure their corporate boards review safety findings.¹⁶ The accord notes that these voluntary practices are not codified into statutory law, though it leaves room for potential future legislation or administrative rulemaking.²³

Industry leaders defended the framework as a workable structure to coordinate safety across commercial rivals. In a public statement following the session, Pichai said Google invested hundreds of billions of dollars across its technology stack over the past two years alone and planned further capital deployment.⁴¹ Pichai said the company focused on pre-release evaluations, red-teaming, and systematic reviews to prevent misuse, describing the accord as a practical basis for shared industry norms.⁴

Zuckerberg emphasized that multi-tiered oversight would help identify flaws before models reach corporate customers and consumers.⁶³ He noted that the agreed commitments establish internal risk assessments coupled with external validation and independent board reviews, calling the accord an opening baseline that could expand across the wider software sector.⁶² Huang said developing defensive safeguards alongside capability advances prevents any inherent trade-off between technical innovation and user protection.¹

The voluntary structure arrives as frontier developers face mounting public concern regarding system stability and physical infrastructure demands. One day before the White House session, OpenAI canceled the public release of its planned model, Astra 6.1, after internal tests indicated the software failed to satisfy internal safety thresholds.³⁵ Independent governance analysts noted that voluntary standards permit companies to set their own risk criteria and lack punitive mechanisms for operational failures.⁵

The gathering also addressed the rapid construction of energy-intensive data centres across the country, where localized power strain and rising utility bills have generated resistance among municipal officials and neighborhood groups.¹³ Trump defended the facilities, saying the participating companies would make substantial contributions to support local infrastructure, schools, and teachers in host regions.¹³ Huang argued that these facilities function as industrial production centres that generate employment and strengthen domestic manufacturing capacity.⁵

Alongside the accord, Trump signed an executive order directing federal departments and agencies to replace the phrase artificial intelligence with the term super intelligence, abbreviated as SI, across official federal correspondence, public websites, and administrative documents.⁵⁶ Trump also signaled plans to appoint a federal official to coordinate policy and discussed forming a 10-member advisory board to monitor development standards.¹²

What this rests on

23 sentences trace to 6 sources.

All 6Documents 1Records 0Reporting 5People 0

1. 1 FI Fortune India Third-party reporting · fortuneindia.com · 30 Sep 2026 Trump, tech leaders sign White House AI accord on super intelligence See the source
2. 2 AR ABC11 Raleigh-Durham Third-party reporting · abcnews.com · 29 Sep 2026 Trump says AI leaders signed a 'constitution' to police themselves See the source
3. 3 "F "@type":"Organization","name":"Agence France-Presse","url":"http://www.afp.com/en" Third-party reporting · afp.com · 29 Sep 2026 Trump touts AI boss pledge to self-regulate See the source
4. 4 H https://x.com/sundarpichai Document · x.com · 30 Sep 2026 AI Companies Sign White House Accord on Super Intelligence After Trump Meeting See the source
5. 5 BN BBC News Third-party reporting · bbc.com · 30 Sep 2026 Three takeaways from Trump's 'Super Intelligence' summit See the source
6. 6 B breitbart.com Third-party reporting · breitbart.com · 30 Sep 2026 Trump, Tech Leaders Sign 'White House Accord on Super Intelligence' See the source