US Lawmakers Target AI Distillation and Propose Tax for Displaced Workers
New legislative efforts aim to sanction foreign firms for 'stealing' AI capabilities and levy taxes on developers to fund worker protections.
US lawmakers are moving to treat artificial intelligence capabilities as strategic national assets and AI revenue as a source for social safety nets. Recent legislative proposals target the practice of AI distillation—where smaller models are trained using the outputs of advanced ones—and seek to tax leading AI developers to mitigate job losses.
In a bipartisan effort, Senators Bill Hagerty (R-TN) and Andy Kim (D-NJ) have proposed an amendment to defense legislation that would blacklist or sanction entities found to be conducting adversarial AI distillation campaigns. The measure specifically targets Chinese companies that use these techniques to improperly extract capabilities from US-developed models. This push follows disclosures from Anthropic to the U.S. Government revealing a massive distillation attack on its Claude model. According to the company, Alibaba executed the campaign between April and June 2026, utilizing approximately 25,000 fraudulent accounts to conduct 28.8 million exchanges.
The Economic Cost of Automation
Parallel to these national security concerns, House Democrats are addressing the domestic economic impact of automation. Representatives Greg Casar (TX), Valerie Foushee (NC), and Sara Jacobs (CA) have championed the "AI Tax and Work Protection Act." This legislation proposes a levy on leading AI companies to fund the creation of a "Work Protection Administration," an agency dedicated to supporting workers displaced by the integration of AI into the workforce.
Strategic and Fiscal Implications
These developments signal a fundamental shift in how the US government views the AI industry. By targeting distillation, the US is effectively treating the internal logic and capabilities of frontier models as intellectual property subject to national security protections. Simultaneously, the proposed AI tax suggests that the financial gains of AI efficiency may be partially reclaimed by the state to offset the social costs of structural unemployment.
If passed, these measures could significantly alter the cost structure for AI developers and heighten geopolitical tensions. The focus on distillation, in particular, creates a new legal and diplomatic friction point regarding how models are trained and the legitimacy of using synthetic data from competitors.
Future Outlook
Industry observers are now watching whether the AI Tax and Work Protection Act gains broader traction in the House or if it remains a partisan proposal. Additionally, the effectiveness of sanctions against distillation remains unconfirmed, as detecting the origin of a distilled model's training data is technically challenging. The coming months will determine if the US can successfully build a regulatory wall around its AI intellectual property while simultaneously funding a transition for its workforce.