Trump's Unacknowledged AI Regulation: Shadow Policy Through Case-by-Case Intervention

The Trump administration came to power vowing to liberate the AI industry from government interference, making deregulation a centerpiece of its technology agenda — scrapping Biden-era AI requirements and publicly arguing that rules would kneecap American innovation. But the reality on the ground looks increasingly different from the rhetoric.

Rather than a hands-off approach, the White House has been quietly shaping the AI industry through targeted, ad-hoc interventions that operate outside any formal regulatory framework. These include overriding state-level AI laws, dictating the national security and cybersecurity review of advanced AI models, steering how AI systems get procured into the federal government, and monitoring the economic footprint of the largest AI companies — all without clear statutory authority or transparent rulemaking.

The tensions are playing out in real time. Anthropic is currently negotiating with the administration over whether export controls on its latest models can be eased, while rival AI labs are scrambling to understand how to comply with Trump's most recent executive order, which created a voluntary framework for government review of certain advanced AI systems. That kind of case-by-case bargaining with individual companies is a form of industrial policy, even if it doesn't carry the label.

The resulting environment is one of significant uncertainty for the industry. Without Congress passing clear national AI rules — which remains its constitutional role — companies face an unpredictable landscape where the rules can shift based on individual negotiations or executive orders rather than stable law. The administration's approach is notably different from structured frameworks like the EU's AI Act or the UK's Online Safety Act, but given that American labs build the world's most powerful models, the practical effect on the global AI industry is arguably larger than any foreign regulation.