AI dangers disturb White House

AI dangers disturb White House

Technology

US administration meets Microsoft, Google CEOs to discuss risks

WASHINGTON (Reuters) - The White House is hosting CEOs of top artificial intelligence companies, including Alphabet Inc's Google (GOOGL.O) and Microsoft (MSFT.O), on Thursday to discuss risks and safeguards as the technology catches the attention of governments and lawmakers globally.

Generative artificial intelligence has become a buzzword this year, with apps such as ChatGPT capturing the public's fancy, sparking a rush among companies to launch similar products they believe will change the nature of work.

Millions of users have begun testing such tools, which supporters say can make medical diagnoses, write screenplays, create legal briefs and debug software, leading to growing concern about how the technology could lead to privacy violations, skew employment decisions, and power scams and misinformation campaigns.

President Joe Biden has also used ChatGPT, a White House official told Reuters. "He's been extensively briefed on ChatGPT and (has) experimented with it," said the official, who asked that they not be named.

Thursday's two-hour meeting which began at 11:45 am ET (1545 GMT), will include Google's Sundar Pichai, Microsoft Corp's Satya Nadella, OpenAI's Sam Altman and Anthropic's Dario Amodei, along with Vice President Kamala Harris and administration officials including Biden's Chief of Staff Jeff Zients, National Security Adviser Jake Sullivan, Director of the National Economic Council Lael Brainard and Secretary of Commerce Gina Raimondo.

"We aim to have a frank discussion about the risks we see in current and near-term AI development," said a senior administration official, speaking on condition of anonymity. "Our North Star here is this idea that if we're going to seize these benefits, we have to start by managing the risks."

Ahead of the meeting, OpenAI's Altman told reporters the White House wants to "get it right."

"It's good to try to get ahead of this," he said when asked if the White House was moving quickly enough on AI regulation. "It's definitely going to be a challenge, but it's one I'm sure we can handle."

The administration also announced a $140 million investment from the National Science Foundation to launch seven new AI research institutes and said the White House's Office of Management and Budget would release policy guidance on the use of AI by the federal government.

Leading AI developers, including Anthropic, Google, Hugging Face, NVIDIA Corp (NVDA.O), OpenAI, and Stability AI, will participate in a public evaluation of their AI systems.

Shortly after Biden announced his reelection bid, the Republican National Committee produced a video featuring a dystopian future during a second Biden term, which was built entirely with AI imagery.

Such political ads are expected to become more common as AI technology proliferates.

United States regulators have fallen short of the tough approach European governments have taken on tech regulation and in crafting strong rules on deep fakes and misinformation.

"We don't see this as a race," the senior official said, adding that the administration is working closely with the U.S.-EU Trade & Technology Council on the issue.

In February, Biden signed an executive order directing federal agencies to eliminate bias in their AI use. The Biden administration has also released an AI Bill of Rights and a risk management framework.

Last week, the Federal Trade Commission and the Department of Justice's Civil Rights Division also said they would use their legal authorities to fight AI-related harm.

Tech giants have vowed many times to combat propaganda around elections, fake news about the COVID-19 vaccines, pornography and child exploitation, and hateful messaging targeting ethnic groups.

But they have been unsuccessful, research and news events show. Just about one in five fake news articles in English on six major social media platforms were tagged as misleading or removed, a recent study by activist NGO Avaaz found.