The White House has proposed new guidelines to broadly shape federal regulations on artificial intelligence, aiming to avoid creating rules that “needlessly hamper AI innovation and growth.”
On Tuesday, the Trump administration released Executive Order 13859, “Maintaining American Leadership in Artificial Intelligence,” which identifies 10 key concerns — including public trust, participation, risk assessment and fairness and non-discrimination.
The order will be introduced Wednesday at the annual Consumer Electronics Show in Las Vegas. It gives federal agencies 180 days to develop a regulatory framework to protect Americans from malicious or intrusive AI systems but also allow tech firms to innovate and thrive.
“AI applications could pose risks to privacy, individual rights, autonomy, and civil liberties that must be carefully assessed and appropriately addressed,” wrote Russell T. Vought, acting director of the Office of Management and Budget.
But Mr. Vought added: “It is not necessary to mitigate every foreseeable risk.”
Critics said the executive order is ambiguous.
“Anything that an agency produces could be shut down, given the vagueness,” Martijn Rasser, senior fellow at the Center for New American Security, told Wired magazine. “It would be easy for anyone to lobby against what’s proposed.”
In response, Lynne Parker, deputy chief technology officer at the White House’s Office of Science and Technology Policy, said the rules are “intentionally high-level.”
“We purposely wanted to avoid top-down, one-size-fits-all, blanket regulations,” Ms. Parker said Tuesday.
Rep. Frank Lucas of Oklahoma, the ranking Republican on the House Science, Space, and Technology Committee, told The Washington Times he welcomed the White House’s AI order.
“The White House recognizes that American competitiveness depends on our ability to develop and use cutting-edge AI technology,” Mr. Lucas said in a written statement. “I appreciate that the Trump Administration is taking a thoughtful, principles-based approach and I’m hopeful that the Science Committee will produce bipartisan legislation this year that furthers this goal and advances American leadership on AI.”
The use of AI has grown rapidly in the last 10 years, but so have concerns about it. A 2018 report from the McKinsey Global Institute estimated that nearly half of all work activities could be automated by AI. Worries about surveillance and facial recognition software led the San Francisco lawmakers to ban the technology in May.
In February, President Trump ordered a “coordinated federal government strategy” on AI.
The 15-page document issued Tuesday addresses a concern that AI risks becoming biased against people of color and women because its developers are mostly white men. For example, Google’s photogenic software in 2015 identified gorillas as “African Americans,” prompting calls that the code was racist.
The document warns agencies that “applications can, in some instances, introduce real-world bias that produces discriminating outcomes or decisions that undermine public trust and confidence in AI.”
Critics also charge the AI can be used maliciously to surveil Americans or be sold to a foreign government to do the same against its citizens. Last week, the Commerce Department announced new restrictions on exports by U.S. companies of AI software “specifically designed to automate the analysis of geospatial imagery.”
The move was widely seen as directed at China, which has been condemned internationally for its use of AI and facial recognition software in the detention of 1 million Uyghur Muslims in reeducation camps in Xinjiang province.
To address such concerns, agencies are directed to draw up regulations being “mindful of any potential safety and security risks, as well as the risk of possible malicious deployment and use of AI applications.”
Please read our comment policy before commenting.