What the election means for AI

AI News

Nov 8

Tuesday’s election saw voters return Donald Trump to the White House and turn over control of the U.S. Senate to Republicans. Control of the House of Representatives still remains too close to call, but Republicans held a slight edge in the final tight races as of Friday afternoon.

So what does this mean for AI and AI policy?

Early reaction from mainstream outlets, the tech press, and AI newsletters is coalescing around a few general expectations. They include:

Trump will quickly repeal President Biden’s Executive Order on AI, which created the first federal guardrails around the emerging technology. Trump is antagonistic toward most government regulation, has promised to kill this particular EO, and the Republican party platform views the Biden plan as one that “hinders AI innovation.” An executive order lives or dies at the pleasure of the President, and this one looks to be a goner.
Don’t look for any extensive Trump-led plan to replace Biden’s AI guidance. Trump leans into destruction, not creation, when it comes to policy and regulation. A fair comp here would be his record on healthcare, which has been heavy on attempts to kill Obamacare and negligible on proposals to replace it. In general, he favors giving capitalism free reign and counting on the market to regulate itself.
Expect far less talk about reducing racial and gender bias in AI, and far more talk about reducing perceived woke strictures and left-wing bias in AI.
When thinking about whether Trump might back or block a particular AI policy, ask this question: Will he view it as helping or hurting the U.S. in an AI arms race against China? Never underestimate the importance of China as an antagonist in the mind of the President-elect.
Elon Musk has the ear of the President-elect this week, and he would seem to be driving White House AI policy. Musk subscribes to the move-fast-break-things philosophy, and is a leading advocate and funder of AI systems that favor anti-woke principles over guardrails limiting hate speech, bullying and misinformation. But remember: Trump advisors come and go with head-spinning frequency. Musk may or may not remain in favor. He may well find himself jettisoned if he’s seen as outshining the man in charge.
The actual work on AI policy could fall to Vice President-elect J.D. Vance. With his Silicon Valley experience and lower public profile, Vance could be the real leader of White House policy on AI, in terms of talking with tech experts and going hands-on with policy creation. He’s skeptical about regulation, and has argued that guardrails would only tighten Big Tech’s grip on the industry.
Two dynamics may clash: Trump’s loathing of business regulations, and his enmity towards Big Tech, which he sees as censoring conservative viewpoints—including, famously, his own. Don’t look for him to give Silicon Valley carte blanche.

Four critical years for AI policy

What happens over the next four years will set the path for AI development and safety regulation for decades to come. “This election could prove the last time the public gets to weigh in on the right way to ‘parent’ AI,” observes Axios tech writer Scott Rosenberg. “Four years from now, the technology will be older — and more set in its ways.”

Congress and AI policy

In Congress, two Republican Senators could rise to power on AI policy. Sen. Ted Cruz (R-Texas), who just won another six-year term, is in line to chair the Senate Commerce, Transportation, and Science Committee, which would exert considerable oversight on AI proposals. Sen. Josh Hawley (R-Missouri) has been working on bipartisan AI proposals and could offer some intra-party pushback against Cruz in this area.

As writer Gopal Ratnam noted in Roll Call this week: “Cruz generally has supported narrow legislation to address harms, such as those from AI-generated pornography and a package of kids’ online safety measures, but he has been very critical of the Biden administration’s approach to regulating AI, as well as bipartisan legislation to create a federal data privacy standard. Cruz has said that too much regulation could choke off innovation and surrender U.S. leadership in the technology to China, calling instead for a hands-off approach that, according to an op-ed he co-authored, would usher in “extraordinary economic growth and prosperity.”

Hawley, by contrast, has been an advocate of tougher regulations on tech companies. He recently worked with Sen. Richard Blumenthal (D-Connecticut) to propose a licensing regime for advanced AI models that would be managed by a federal agency.

Gabby Miller and Justin Hendrix at Tech Policy Press noted that the most progress on AI issues was made in this past Congress through bipartisan participation. Will that continue? They wrote:

“It is unclear whether any of the goals of the Bipartisan Senate AI Working Group roadmap released following the 2023 AI Insight Forums will survive in a Republican-led Senate. The forums were led by Senate Majority Leader Chuck Schumer (D-NY) in partnership with Sens. Mike Rounds (R-SD), Martin Heinrich (D-NM), and Todd Young (R-IN). Civil society groups widely criticized the roadmap they produced for prioritizing industry interests and foregrounding competition with China, but Republicans may push things in the other direction.”

state legislators will have increased power

Meanwhile, most of the real action in AI policy has been happening in state legislatures. Look for that to continue. In fact, state legislators may now have more power than ever, in terms of creating the rules of the road for AI safety and transparency. Congress has failed to pass any substantial legislation on artificial intelligence, and the Republican takeover of the White House and (possibly) both houses of Congress makes the emergence of a major AI bill even less likely.

At the local level, 50 state legislatures will open their next sessions in January 2025, weeks before President-elect Trump is sworn into office. A second Trump White House may speak loudly on AI policy, but those state legislators could set the actual agenda for this quickly evolving technology.