Top Tech CEOs Discuss A.I. Safety at the White House: What They Agreed Upon

Seven leading A.I. companies agreed to open up their systems for public evaluation.

Sam Altman wearing a gray jacket.
OpenAI CEO Sam Altman Matt Winkelmeyer/Getty Images for WIRED25

The CEOs of Google, Microsoft (MSFT), OpenAI and Anthropic—four leading companies in artificial intelligence—were invited to the White House today (May 4) to meet with Vice President Kamala Harris and senior officials of the Biden administration and discuss the safety of A.I., a hotly debated topic in the tech community in recent weeks.

Ahead of the meeting, the White House announced three new initiatives, including $140 million in federal funding for A.I. research, a commitment by leading A.I. firms to open up their models for public evaluation, and an upcoming policy draft about how the U.S. government will use A.I.

Tech CEOs arrive in Washington, D.C.

Invited to today’s meeting were Google CEO Sundar Pichai, Microsoft’s Satya Nadella, OpenAI’s Sam Altman, and Anthropic cofounder and CEO Dario Amodei.

Altman and Pichai were spotted outside the White House this morning speaking to reporters. Altman told a group of reporters ahead of his meeting that A.I. is “definitely going to be a challenge” for society but one that “we can handle,” according to videos posted on Twitter. Pichai said A.I. is “an important topic” and he looks forward to a productive meeting, according to a tweet by Reuters reporter David Shepardson.

At around 3 p.m. today, Pichai was seen leaving the White House. He declined to comment on his meeting with Harris, Shepardson reported.

Top A.I. firms agree to open up algorithms for public review

Seven tech companies—Google, Microsoft, OpenAI, Anthropic, Stability AI, Nvidia, and Hugging Face—have agreed to open up their A.I. systems for public evaluation on a platform developed by Scale AI, a company that provides data for training machine learning algorithms.

This process is to ensure these companies’ models align with principles and practices outlined in the Biden administration’s Blueprint for an A.I. Bill of Rights and A.I. Risk Management Framework, the two documents introduced by the Biden administration in the past six months to promote responsible innovation.

“This independent exercise will provide critical information to researchers and the public about the impacts of these models, and will enable AI companies and developers to take steps to fix issues found in those models,” the White House said in a statement.

While the A.I. race is largely led by tech behemoths like Google and Microsoft (also an investor in OpenAI), startups specializing in large language models are rapidly gaining traction. Anthropic, Stability AI, and Hugging Face are among the leading players in the A.I. startup space. Anthropic, a barely two-year-old company, was recently valued at more than $4 billion. 

Biden administration commits funding and policy guidance

The National Science Foundation, a federal agency that supports fundamental research and education, will spend $140 million establishing seven new National A.I. Research Institutes to help drive responsible A.I. innovation, the White House said today.

The federal government currently runs 18 such A.I. research institutes across the country. The new ones will focus on A.I. research in areas deemed as critical by the government, including climate, agriculture, energy, public health, education and cybersecurity.

In addition, the Biden administration’s Office of Management and Budget will release draft policy guidance this summer on the use of A.I. systems by the U.S. government. The guidance, which will be open for public comment, aims at establishing rules that ensure government agencies’ development, procurement, and use of A.I. systems align with the American people’s rights and safety. Top Tech CEOs Discuss A.I. Safety at the White House: What They Agreed Upon