When the Trump administration takes control in January, many observers expect the Biden administration's executive order on AI, which imposed some standards and protections around the development and use of the technology, will be dismantled. What that order will be replaced with is uncertain and to some extent subject to the whims of influential advisors like Elon Musk and Peter Thiel.
The Trump-led government is bound to be pro-business and anti-regulation, as it was the last time.
"I assume the policy will focus on American businesses and workers to maintain global competitiveness in AI development," said Ryan Hildebrand, chief innovation officer at Bankwell, a Connecticut community bank based in New Canaan with just over $3 billion of assets. "It will aim to protect American jobs and intellectual property, possibly through restrictions on foreign AI companies and favoritism for U.S.-based developers."
But this does not mean banks can relax. Bank regulators are likely to continue to require them to comply with existing laws, including antidiscrimination and fair credit rules, in their use of AI.
What a Trump AI policy might include
Everyone interviewed for this article agreed the
"Trump has stated that he will repeal Biden's executive order on AI," said Alenka Grealish, principal analyst at Celent.
The Trump administration will try to put its own stamp on AI policy, said Aaron Cummings, a partner at Crowell & Moring.
"I would characterize President Biden's approach as very cautious and very heavy on making assignments to different U.S. agencies to develop a lot of regulations, because of the great concerns about how disruptive this could be to society and some of the dangers that are involved," Cummings said. "The Trump administration is very conscious of all of those things, but they also believe that we have an opportunity to establish an artificial intelligence industry where the United States can be a world leader. So the Trump administration wants to make sure that it doesn't crush this nascent industry with a heavy regulatory hand."
The
"The AFPI has a lot of really sharp people and they are undoubtedly working on this issue," Cummings said. "They've published a lot on this issue already, and so I wouldn't be surprised at all that they're taking a lead role in preparing various policy documents that the Trump administration will use."
In contrast to Biden's executive order, which emphasizes global collaboration and equity, "an 'America First' policy would likely be more protectionist, focusing on national security and American economic dominance in the AI sector," Hildebrand said.
But also influential are close Republican advisors like Musk, Thiel and J.D. Vance.
"I think there's no doubt that Elon Musk will have as big a role as he might like to play in Trump's thinking on how to approach AI policy," Cummings said. "He clearly has Trump's ear, and he's also an expert in the field. I think he is going to be one of the most influential voices on AI policy."
Gilles Ubaghs, strategic advisor at Datos Insights, pointed out that there's a paradox in the way Trump and his advisors look at regulation.
"While Republicans have historically and continue to espouse deregulation, the populist tide of recent years is generally quite negative to big tech with lots of talk on the perceived unfairness of social media," he said. "Add to that the arguably newer strain of economic populism from J.D. Vance, and there's a push and pull here on bringing big tech to heel while also wanting to deregulate. Trump tends to go with whoever had his ear last, and so now you've got people like Peter Theil with his backing of Vance, and now Elon Musk doing whatever it is he's doing, so their influence is likely to be really decisive at least for the time being. Musk and big tech generally don't like regulations of any kind."
On the other hand, Musk is also part of the effective altruism movement, which worries about the long-term dangers of AI and the threat it poses to humanity.
"The effective altruists in theory are at least publicly cautious about AI," Ubaghs said. "Then again X.AI is a very real investment so again these are contradictory views."
Elon Musk's involvement makes it "difficult to predict how the Trump administration will regulate AI, if it does at all," Grealish said. "The prediction challenge lies in the fact that Elon Musk has become a powerful influencer and his views could differ from Republicans."
Taking all this into account, it's most likely that the Biden AI order will be rescinded, but something probably very similar will take its place, Ubaghs said.
"Some of the current guidelines on forming industry standards are likely to be jettisoned, but some sort of similar voluntary code will likely emerge in the near future," he said. "What is fully off the table for the foreseeable future is any sort of regulatory mandate with real bite. Voluntary codes of practice are good and a nice starting point – but realistically without stricter regulatory regimes in place, there's nothing to really stop technologists doing what they want."
There's likely to be a focus on industry self regulation, Ubaghs said.
No free passes for banks
The coming change in administration and AI policy does not mean banks should be less cautious with AI deployments, Hildebrand said.
"On the contrary, there is a need to invest more in data security, ensure AI models are well understood, comply with evolving regulations and mitigate reputational risks," he said. "The worst outcome would be a significant incident related to AI use, which could erode customer trust in banking. Such an event would severely hinder innovation, regardless of policy and politics."
Given the rapidly evolving AI landscape, banks need to prioritize the safety and privacy of customer data, ensure AI is unbiased, communicate clearly about how AI functions, closely monitor policy changes and train staff to use AI responsibly, Hildebrand said.
"Regardless of regulatory actions, banks should continue to focus on fulfilling first principles, such as ethical use, accountability, and transparency," Grealish said.
Banks still need to contend with existing regulations and rules, none of which have changed, Ubaghs said. They also have to be wary of the liability and reputational risk of working with sketchy AI tools.
"The potential for fraudulent and grifting AI solutions is likely to increase with a less regulated environment," he said. "Most banks remain extremely cautious about their AI projects, and they should carry on."
If anything their jobs now become harder as banks will need to double down on due diligence for these solutions, he said.
"Ironically, greater regulation and standardization would likely help the market enormously by creating a baseline for the technology and alleviate many of the fears and reasonable concerns that the market across almost every industry now holds," Ubaghs said.
Cummings advises industry leaders to speak up and become part of the conversation.
"There is such a huge appetite for policymakers to hear from companies that are trying to deploy this technology in effective ways and who are grappling with all of the issues that are being discussed," he said. "And I think there's a real opportunity for companies that are already using this to help policymakers understand the risks and the opportunities that are available and to help safeguard some of the innovation that is happening now."