Executives from Google, Anthropic, Meta, OpenAI, xAI, and Nvidia signed the White House Accord on Super Intelligence on Tuesday following a White House luncheon.
The voluntary agreement calls for internal model monitoring, dedicated compliance teams, outside independent audits, and board-level oversight committees.
Donald Trump announced at the UN General Assembly on 22 September 2026 that US federal agencies will replace the term artificial intelligence with Super Intelligence.
Anthropic testing revealed Chinese open-weight model GLM-5.3 succeeded in 50 out of 410 cyber exploit tests, compared with 56 successes for Claude Mythos Preview.
On Tuesday, President Donald Trump hosted tech industry leaders at a White House luncheon to sign The White House Accord on Super Intelligence. Six technology companies signed the voluntary agreement: Google, Anthropic, Meta, OpenAI, xAI, and Nvidia. Meta chief executive Mark Zuckerberg served as the primary architect of the pact, collaborating with Nvidia chief executive Jensen Huang to pitch the framework to senior administration officials and tech leaders. House Speaker Mike Johnson and Amazon executive chairman Jeff Bezos also attended the meeting.
Trump framed the pact as an act of self-regulation by technology firms. Speaking outside the White House alongside company executives, Trump described the commitments as morally binding.
“They're really going to be policing each other.”
The signing ceremony followed a campaign by Trump to replace the phrase artificial intelligence with Super Intelligence across federal agencies. Trump argued on Truth Social that the term artificial intelligence was inexact and unappealing. In a three-day survey conducted on the platform, 65% of participants favored the term Super Intelligence over Superior Intelligence, which received 35% of votes.
Data from the article text
Trump reiterated the terminology shift during an address at the United Nations General Assembly on 22 September 2026, stating that future American federal documents would adopt the name. Computer science researchers questioned the rebranding. Pedro Domingos, a computer science professor at the University of Washington, pointed out that existing AI models do not match the standard definition of superintelligence, which refers to software surpassing human abilities across all fields.
“Most of AI is not superintelligence. It is just a different type of intelligence.”
The accord outlines four internal governance commitments for the participating companies. Signatories agreed to maintain internal controls monitoring model capabilities and alignment during training and deployment, empower internal compliance teams, work with independent external auditors, and establish dedicated oversight committees on their corporate boards. The pact contains no legal enforcement mechanisms or mandatory penalties for non-compliance, nor does it mandate the public disclosure of audit findings.
Following the event, Google chief executive Sundar Pichai stated on social media that the agreement provides a foundation to build public trust. Nvidia chief executive Jensen Huang dismissed suggestions that safety measures would hinder commercial progress.
“There's no conflict between innovation, technology and safety.”
The voluntary agreement mirrors commitments signed during the Biden administration in July 2023, as well as the 2025 Frontier AI Safety Commitments signed in the United Kingdom and South Korea. However, the agreement arrives alongside internal industry debates regarding model safety. OpenAI delayed the deployment of its GPT-6.1 Astra model after safety systems lead Saachi Jain confirmed the system fell short of safety thresholds.
Competition with foreign labs remains a consideration for the signatories. A report from Anthropic revealed that GLM-5.3, an open-weight model created by Chinese firm Z.ai, successfully completed 50 out of 410 automated cyber exploit tests, compared to 56 successful attempts by Anthropic's Claude Mythos Preview.
Data from the article text
While companies signed the voluntary compact, the Federal Trade Commission is preparing inquiries into multiple artificial intelligence firms regarding potential consumer protection concerns.
Pollar is free and carries no ads in the text. Its members keep it going.
Support PollarAI & Tech · Washington
Six leading tech firms signed a voluntary safety agreement at the White House on Tuesday, committing to internal governance measures as Donald Trump pushes federal agencies to rename artificial intelligence to Super Intelligence.
AI text based on reports by Deutsche Welle, POLITICO, Washington Post and other publishers. Changed .