The White House will take wide-ranging action on Monday to increase artificial intelligence (AI) safety while protecting consumers, workers, and minority groups from the technology’s related risks. The executive order President Joe Biden will unveil is the latest step by the administration to set parameters around AI as it makes rapid gains in capability and popularity. It comes as the EU puts the finishing touches on a set of regulations that will target the riskiest applications for AI, and other countries are working to establish their guidelines.
The to-do list in the executive order, which includes new security and safety standards for AI, focuses on privacy, civil rights, consumer protections, scientific research, and worker rights. The order also looks to enhance U.S. leadership in AI through advanced research and funding for businesses that are using the technology to advance society, with a particular focus on small companies.
Many of the new initiatives in the executive order will be implemented over time, with some launching within 90 days and others requiring between 9 to 12 months to complete, an administration official who previewed it on Sunday told reporters. The broader measures will aim to create AI transparency and accountability, including establishing a cybersecurity program that can use AI to find flaws in critical software infrastructure, creating an initiative that helps develop AI talent, and easing visa requirements for overseas tech workers who are crucial to the development of the sector.
Some of the more significant measures include a requirement for developers of the most powerful AI systems to share their safety tests with the government and invokes the Korean War-era Defense Production Act to track companies developing “frontier” models, or advanced versions of existing technology that could impact national security, economic security or national public health and safety. Other parts of the order aim to address discrimination in the workplace by setting guidance for landlords, federal benefits programs, and federal contractors and coordinating with Department of Justice civil rights offices to ensure that AI tools don’t exacerbate bias. It also seeks to boost upskilling efforts in the workforce to prepare for significant job displacement resulting from AI and requires a report on labor market impacts from the technology.
Other components of the executive order will include a directive to the Commerce Department to create guidance for AI watermarking and an effort to use AI to identify fake news and propaganda. The measure will also urge the Federal Communications Commission to use AI to block unwanted robocalls and texts, as well as create an “AI talent” task force to ramp up hiring efforts for federal agencies. It will also ask Congress for additional legislative authority to further enhance AI governance. The executive order will have the force of law, but the administration believes it needs to be bolstered by a broader package of legislation to be fully effective. That’s something the Democratic Senate majority leader, Chuck Schumer of New York, has pushed for and is expected to introduce this year.