Artificial intelligence represents an unprecedented economic opportunity as well as a complex security challenge for the United States. As Trump gets set to begin his second term, he should prioritize a comprehensive AI agenda that spans deregulation, promoting domestic energy production, and enacting infrastructure permitting reform. This broad-based approach will ensure continued American leadership in AI.
Repealing Biden’s AI Executive Order
The Biden administration’s more than 100-page AI executive order should be repealed on Day 1 of Trump’s term. Though certain elements of the order are reasonable and mirror Trump’s more focused 2019 AI order—such as leveraging AI to improve government operations—the order’s sweeping regulatory framework demands immediate revision.
At its core, Biden’s directive imposes burdensome reporting requirements on private companies based on the results of safety assessments of AI systems. While rigorous analysis of this kind might suit government operations well—where, absent strict analytical requirements, there is often little incentive to take account of the best evidence or science—it is ill-suited to the private sector’s dynamics. Over time, these assessments will evaluate an exhaustive list of potential risks, creating a bureaucratic maze that will slow development cycles.
Companies already have compelling incentives to ensure their AI products do not harm their customers. Their reputations and bottom lines depend on it. They are typically the ones best positioned to assess the degree of pre-market testing their models require, given the unique tradeoffs each firm faces with respect to risk, competitive pressures, customer needs, and shifting market dynamics. Government regulators, by contrast, are ill-equipped to gauge what is safe and what isn’t in a market that is evolving as rapidly as AI.
In spite of these lessons, the National Institute of Standards and Technology is making moves to codify impact assessment procedures into its formal guidance documents. The Biden administration’s regulatory approach thus mimics failed models seen in California’s rejected AI legislation and the European Union’s restrictive AI Act. America cannot afford to follow Europe’s innovation-killing regulatory model that cedes technological leadership to competitors like China.
Reframing AI Risk as a National Security Imperative
AI risk should be approached as a critical national security challenge comparable to counterterrorism. The most significant threats will come from decentralized malicious actors and hostile states weaponizing AI technology to achieve widespread harm. Like counterterrorism policy, which needs real-time intelligence and flexible security responses (like quickly updating no-fly lists or adapting airport screening procedures), AI oversight must be agile enough to address emerging threats (such as new types of AI-enabled fraud or cyber attacks).
While the Biden administration has made some progress on cybersecurity—establishing a national cybersecurity strategy that includes some reasonable provisions—its overall approach to AI safety through top-down bureaucratic rules remains too rigid.
California’s rejected SB 1047 demonstrates the pitfalls of the bureaucratic approach. It positions big tech companies as adversaries in the AI safety effort. These companies are not the source of our national security threats, however; instead, they will help serve as our first line of defense against foreign cyberattacks and AI misuse by hostile states or bad actors.
To the extent possible, Congress should preempt such misplaced state-level initiatives, establishing instead a coherent national framework that recognizes the strategic and national security benefits of a strong U.S. tech sector.
Aligning Energy and AI Policy
The United States needs reliable and abundant energy across the board. By building a resilient electricity grid that meets this standard, AI systems will have the energy they need to stay competitive with foreign competition, including from China.
While many tech companies advocate for renewable energy solutions to address AI energy needs, these options aren’t always practical or economically viable. The current practice of supporting net zero and similar climate initiatives through renewable energy subsidies should be abandoned. These policies distort energy markets, funnel taxpayer dollars to wealthy corporations, and promote a woke corporate agenda at the expense of the energy needs of the public.
Repealing Biden’s Inflation Reduction Act subsidies should therefore be a top priority. To the extent conflicts arise between tech companies’ climate goals and AI energy needs, meeting the nation’s computing power requirements should take precedence.
Similarly, transmission line reforms that aim to pass off the costs of a renewable energy “transition” to utility rate payers constitute a step in the wrong direction. While winning the AI race is important, the needs of the tech industry don’t necessitate public subsidies that hurt consumers. As one of the largest sectors of the American economy, the tech industry is perfectly able to stand on its own two feet without government support.
The Trump administration can still promote practical ways to bring more energy production online, and this will benefit both the tech industry and the public more generally. Congress’s recently passed ADVANCE Act redirects the Nuclear Regulatory Commission’s mission toward supporting domestic nuclear development. To build on that effort, overly-conservative radiation standards that drive up the cost of nuclear power plants should be revisited.
Similarly, the Biden administration’s Power Plant Rule, which seeks to phase out large swaths of the domestic coal and gas fleet, should be repealed as it constitutes an attack on the energy production the U.S. needs for a thriving economy.
Co-locating data centers with new or existing power plants—an approach that has faced some resistance from the Federal Energy Regulatory Commission (FERC)—is another wise approach. Trump’s FERC should avoid standing in the way of these beneficial partnerships.
Streamlining Permitting for AI Infrastructure
The Trump administration should prioritize overhauling National Environmental Policy Act regulations related to permitting of critical infrastructure, such as data centers and new power sources. A recent court decision has limited White House authority over setting environmental review standards. Even if this ruling holds up, individual agencies still maintain their own review requirements that supplement White House regulations. These regulations need to be revised and reconsidered.
Updates should include raising the thresholds for actions requiring environmental review. This can be achieved by clarifying the definition of “major Federal action”—to mean the $100 million impact standard defined by Congress—and establishing clear, measurable criteria for what constitutes actions “significantly affecting the quality of the human environment.”
Charting a Strategic Course
President Trump’s second term offers a once in a lifetime opportunity to shape AI governance for Americans’ benefit. By turning away from the top-down bureaucratic model that defined Biden’s approach to AI governance, the Trump administration can support American innovation while still protecting vital national security interests. Embracing a nimble, adaptive AI strategy will keep Americans safe while preserving the values and competitive advantages that have made the U.S. the world’s leading innovation hub.