Earlier this week, Trump rescinded executive order (EO) 14110, which established risk management guidance for the technologies. That order dates to October 2023, and some experts argued it did not go far enough to regulate AI. Trump’s decision to revoke that order is one that was expected prior to his taking office. At that time, experts said they expected the new administration’s AI strategy to focus less on mitigating discrimination and more on bolstering the workforce and cybersecurity.
UNPACKING THE NEW EO
The EO requires development of AI systems in the country to be “free from ideological bias or engineered social agendas,” although the specifics of what this means are not made clear in the order.
It also calls for the creation of an AI Action Plan, within the next 180 days, to support “AI dominance”; the plan’s creation is to be led by the assistants to the president for science and technology, and for national security affairs, and by the special adviser for AI and crypto.
Furthermore, the new EO directs the White House to modify and reissue the Office of Management and Budget’s memoranda — M-24-10 and M-24-18 — on AI within 60 days. These memoranda provide guidance for federal government departments and agencies on risk management, acquisition and governance of the technologies; the aim is to eliminate barriers hindering the country’s leadership in AI.
A fact sheet on the new EO cites past action under Trump on AI, including his 2019 EO on AI and guidance for federal agency adoption established in 2020. According to the fact sheet, the new EO builds on those efforts to support U.S. leadership in AI, “rooted in free speech and human flourishing.” The order makes clear that nothing within it should affect “the authority granted by law to an executive department or agency.”
When asked whether he thinks that AI will replace U.S. jobs, Trump definitively responded after signing the AI EO this week that it will not, but rather that it will create “tremendous numbers of” jobs.
This appears to align with what Aspen Digital Senior Director Zaki Barzinji said prior to Trump’s EO, underlining his belief that Trump’s AI work would prioritize the bipartisan goal of futureproofing the workforce.
INITIAL STAKEHOLDER REACTIONS
Eric Gastfriend, executive director of Americans for Responsible Innovation — a nonprofit organization working on public interest policy advocacy — said in a statement that repealing the 2023 EO this week has already led federal agencies to freeze work on AI policies, arguing that the new EO “shouldn’t come as a surprise.”
“Today’s executive order is a placeholder until the administration has a chance to develop a full strategy for executing that vision,” Gastfriend said.
Quinn Anex-Ries, senior policy analyst on the Center for Democracy and Technology’s Equity in Civic Technology team, said of federal actions like AI inventories that states see the success of these initiatives at the federal level, and then in turn decide to implement state-specific versions.
“I think what we’ve already seen is that states are learning from different approaches at the federal level,” Anex-Ries said.
Following the latest action this week, Anex-Ries said via email that the guidance documents that were issued under the October 2023 executive order can still be a helpful resource for state and local governments to implement AI governance.
“Regardless of what happens at the federal level, state and local governments have an obligation to ensure that public agencies’ use of AI is transparent, safe, and aligned with existing legal obligations,” Anex-Ries said.
Prior to the inauguration, American Society for AI Chair Aaron Poynton said he thought eliminating the 2023 EO would be “a good start.” In an email following the new EO, he said it is “a step in the right direction.”
“American companies need to be focused on innovation to ensure American leadership,” Poynton wrote.
