New Rules Coming for Use of AI and Chat AI by Government

By Courtney Rozen and John Hewitt Jones
Bloomberg News- (via TNS).

President Joe Biden will sign a wide-ranging directive on artificial intelligence on Monday, according to people familiar with the matter, as part of a broader pledge to steer AI models away from potential misuse before they’re released to the public.

The White House will also be hosting an event on “safe, secure and trustworthy” artificial intelligence Monday, according to the people, who spoke on condition of anonymity to discuss the information. Some executives from the tech industry will be in attendance.

The executive order is expected to include new policies such as requiring AI models to undergo assessments before they can be used by federal agencies and easing visa requirements for overseas technology workers, the Washington Post reported earlier Wednesday.

It is also expected to mandate assessments of government-purchased large language models — the technology that powers AI chatbots — through “red teaming” exercises undertaken by the National Institute of Standards and Technology, a laboratory operated by the Department of Commerce that sets technology standards for government and the private sector, according to the report.

Further, the order could require government agencies including the Defense Department, the Energy Department and intelligence agencies to run assessments to determine how they may incorporate AI into their work.

The White House is slated to issue the order days before Vice President Kamala Harris and industry leaders attend a summit in the UK about AI risks, led by Prime Minister Rishi Sunak. The executive order could give the Biden administration a piece of AI policy to point to on the world stage at a time when the European Union and China are further along in developing AI regulations.

Biden has spoken before about the threats of artificial intelligence, and has promised to protect consumers while calling on leading AI companies to meet voluntary transparency and security standards.

“We must be clear-eyed and vigilant about the threats emerging technologies that can pose — don’t have to, but can pose — to our democracy and our values,” Biden said in July.

The U.S. Congress is separately debating whether and how to impose binding regulation on AI. Senate Majority Leader Chuck Schumer, a Democrat from New York, has repeatedly stressed the need to create guardrails from risks like disinformation and bias, while promoting its potential to spur advancements across fields, such as medicine, education and finance.

Schumer has hosted tech executives and investors as part of closed-door meetings with Congress in recent weeks, including Meta Platforms Inc. Chief Executive Officer Mark Zuckerberg, Microsoft Corp.’s Satya Nadella, Tesla Inc.’s Elon Musk, venture capitalists Marc Andreessen and John Doerr, and OpenAI’s Sam Altman.

_____

(With assistance from Anna Edgerton, Oma Seddiq and Justin Sink.)

_____