Back to News
Market Impact: 0.6

Don’t be fooled. The US is regulating AI – just not the way you think | Sacha Alanoca and Maroussia Lévesque

Artificial IntelligenceRegulation & LegislationTechnology & InnovationGeopolitics & WarSanctions & Export ControlsTrade Policy & Supply ChainElections & Domestic Politics
Don’t be fooled. The US is regulating AI – just not the way you think | Sacha Alanoca and Maroussia Lévesque

Contrary to widespread perception, the US government is not deregulating artificial intelligence but is instead strategically intervening to control its foundational 'building blocks,' such as advanced chips and model weights, primarily for national security reasons. This approach, often obscured by complex administrative language, marks a significant shift from earlier application-focused regulations (e.g., EU) towards an 'iron grip at the core' of the AI technology stack. This redefinition of where regulation occurs, focusing on underlying infrastructure rather than visible applications, has profound implications for the AI industry, global competition, and the future of international AI governance.

Analysis

The prevailing narrative of US AI policy as deregulatory is a misconception; instead, the US government is strategically intervening with an "iron grip" on the foundational "building blocks" of AI, such as advanced chips and model weights. This hands-on approach, exemplified by both the Trump and Biden administrations, prioritizes national security concerns over a laissez-faire stance on the technology's core infrastructure. Biden's restrictions on chip access to China and the control over model weights underscore this strategic intervention. This US strategy represents a significant shift in global AI regulation, moving from application-focused rules (like the EU's AI Act addressing societal harms) to an emphasis on underlying components. The "second wave" of regulation, led by the US and China, is driven by national security, military advantage, and preventing malicious actors from leveraging AI. This contrasts with earlier frameworks that targeted visible applications and concerns like discrimination or surveillance. The article highlights that Washington's interventions, often buried in complex administrative language like "Implementation of Additional Export Controls," are a clear trend of regulating where most people aren't looking. This redefinition of regulatory focus implies that companies involved in AI hardware, data centers, and core software components face increasing governmental oversight and potential restrictions. The authors argue for greater transparency regarding these strategic interventions, challenging the "deregulation myth."