US President Donald Trump and the chief executives of six of the world's largest technology companies have signed a voluntary White House accord laying out new safety and oversight measures for advanced artificial intelligence systems.
The signing took place on Tuesday at the White House during a luncheon that brought together leaders from Anthropic, OpenAI, Google, Meta, xAI and Nvidia. The executives present were Dario Amodei, Greg Brockman, Sundar Pichai, Mark Zuckerberg, Elon Musk and Jensen Huang.
**Four layers of oversight**
Trump released the "White House Accord on Super Intelligence" on his Truth Social platform, while the signed text was titled the "Joint Commitment on Frontier Responsibilities". It calls for four defensive tiers of internal controls:
- Continuous monitoring of advanced models during training, targeting high-risk capabilities such as biosecurity threats, chemical weapons synthesis and cyber intrusions
- Internal oversight groups inside each company responsible for auditing controls and addressing vulnerabilities in real time
- Independent external evaluators to periodically stress-test safeguards and publish performance reviews
- An independent committee within each company's board that receives auditor findings directly, bypassing executive leadership
The companies also pledged to meet regularly to establish common standards and best practices for improving the safety of their systems.
**Voluntary, for now**
Trump described the agreement as "morally binding" and said the tech leaders had effectively agreed to check one another's work. The accord leaves open the possibility that the measures could eventually be codified into law or formal regulations.
The signing came alongside a September 29 executive order directing the US federal government to use the term "superintelligence" instead of "artificial intelligence" in official documents where legally possible. Trump has consistently favoured industry self-regulation over heavier government oversight, arguing that the US must not cede its AI edge to China.
Sources:
- Al Jazeera: Trump, tech bosses sign voluntary pact pledging 'robust' AI safeguards
- The Indian Express: Trump, tech leaders sign 4-tier pact for AI self-regulation
- Business Standard: US AI accord: Tech companies to police AI amid calls for global regulations
The signing took place on Tuesday at the White House during a luncheon that brought together leaders from Anthropic, OpenAI, Google, Meta, xAI and Nvidia. The executives present were Dario Amodei, Greg Brockman, Sundar Pichai, Mark Zuckerberg, Elon Musk and Jensen Huang.
**Four layers of oversight**
Trump released the "White House Accord on Super Intelligence" on his Truth Social platform, while the signed text was titled the "Joint Commitment on Frontier Responsibilities". It calls for four defensive tiers of internal controls:
- Continuous monitoring of advanced models during training, targeting high-risk capabilities such as biosecurity threats, chemical weapons synthesis and cyber intrusions
- Internal oversight groups inside each company responsible for auditing controls and addressing vulnerabilities in real time
- Independent external evaluators to periodically stress-test safeguards and publish performance reviews
- An independent committee within each company's board that receives auditor findings directly, bypassing executive leadership
The companies also pledged to meet regularly to establish common standards and best practices for improving the safety of their systems.
**Voluntary, for now**
Trump described the agreement as "morally binding" and said the tech leaders had effectively agreed to check one another's work. The accord leaves open the possibility that the measures could eventually be codified into law or formal regulations.
The signing came alongside a September 29 executive order directing the US federal government to use the term "superintelligence" instead of "artificial intelligence" in official documents where legally possible. Trump has consistently favoured industry self-regulation over heavier government oversight, arguing that the US must not cede its AI edge to China.
Sources:
- Al Jazeera: Trump, tech bosses sign voluntary pact pledging 'robust' AI safeguards
- The Indian Express: Trump, tech leaders sign 4-tier pact for AI self-regulation
- Business Standard: US AI accord: Tech companies to police AI amid calls for global regulations
