Trump Signs Voluntary AI Safety Agreement With Tech Giants, Backs U.S. Data Center Expansion

U.S. President Donald Trump announced on Tuesday after meeting technology executives at the White House that major technology companies had agreed to establish voluntary industry standards for artificial intelligence and support the rapid expansion of data centers in the United States. The agreement requires companies to engage independent auditors to assess whether AI systems operate as intended. Companies will also make best efforts to prevent their own tools from gaining unauthorized access to or intruding into technical systems in unintended ways.

Previously, both OpenAI and Anthropic reported that their AI agents had escaped their control and intruded into other companies’ systems, drawing greater scrutiny of the risks posed by advanced AI. The expansion of data centers has also faced community resistance, with some areas opposing construction because of concerns about electricity consumption, public utility expenses and other issues. This has placed the Republican Party under political pressure ahead of the November 3 midterm elections.

Executives from OpenAI, Anthropic, Meta, Google and Nvidia all attended. Following a closed-door lunch, Trump, House Speaker Mike Johnson and the chief executives appeared at a public event on the White House driveway and answered questions from reporters for approximately 30 minutes. Meta CEO Mark Zuckerberg said the companies would establish “robust internal controls” for AI systems.

Trump compared the agreement to a “charter,” saying that some of the world’s largest companies had signed it and that he had also signed it in his capacity as president. He is considering the creation of a 10-member committee to oversee AI safety and plans to appoint a new official to coordinate White House policy. The potential members of the committee have not been announced.

When questioned about the possibility that data center construction could affect Republican electoral prospects, Trump said technology companies would promote community acceptance of the projects and bring substantial investment. However, a Reuters and Ipsos poll conducted from September 17 to 20 found that 73% of Americans were concerned that companies would fail to do enough to prevent AI from causing serious harm, while 55% believed that its development should be slowed.

Also on Tuesday, Trump’s aides unveiled a new AI tool designed to make it easier for the public to use government services, with the project led by Airbnb co-founder Joe Gebbia. Trump said the tool theoretically could not be hacked by bad actors, but he provided no evidence to support that claim. He also directed government agencies to refer to the technology as “superintelligence” rather than “artificial intelligence.”

Turn any website into clean JSON.

Ready-made scrapers & AI agents

Try Apify