Tyler Cowen argued that artificial intelligence could remain lightly regulated for now because future rules may become excessive, a forecast about the timing of oversight that he did not tie to a specific U.S. policy proposal.
In an Oct. 22 post, Cowen said his view reflects a broader belief that much of the American economy becomes overregulated once rules are imposed. He offered no data to support that claim and did not define what level of oversight would count as too much or too little. “After all, overregulation will have decades, maybe centuries, to catch up and even the ledger,” he wrote.
The European Union’s Artificial Intelligence Act is being implemented in stages. The European Commission says enforcement of certain provisions began Aug. 2, 2026. They include transparency requirements for some systems, such as telling users when they are interacting with AI and labeling deepfakes.
Under the revised timetable, rules for certain high-risk AI systems apply from Dec. 2, 2027. Requirements for high-risk AI embedded in regulated products apply from Aug. 2, 2028. The Commission says the EU’s Digital Omnibus on AI entered into force July 27, 2026, extending some implementation deadlines.
Those dates set out the EU’s rollout. Cowen’s post did not cite a pending U.S. action that would test his forecast, or explain how policymakers should judge whether current AI oversight is insufficient. The post also did not identify who might benefit from delaying rules or who might bear the costs.
