TL;DR: Amodei proposed taxing AI companies to fund UBI on June 10-11. Sanders introduced a sovereign wealth fund bill on June 18. Neither is law. Both signal that AI-driven job displacement is being framed as a corporate externality, and that disclosure or reporting requirements are coming.
Two things happened in the same week that rarely happen in AI policy discussions: the CEO of a leading AI company proposed taxing his own industry, and a U.S. senator introduced legislation to do exactly that.
On June 10-11, 2026, Dario Amodei published an essay proposing that governments should tax AI companies to fund universal basic income for workers displaced by automation. Anthropic simultaneously announced a $200 million research initiative to study AI's actual impact on the labor market. On June 18, the same week, Senator Bernie Sanders introduced the first major standalone federal bill to tax AI firms and channel the proceeds to a sovereign wealth fund for displaced workers.
Neither proposal is law. Neither will become law in its current form quickly. But together they mark a shift: AI-driven job displacement is being reframed from a speculative long-term risk into a present economic externality, and policymakers and industry leaders are publicly arguing about who should pay for it.
For enterprise governance teams, the timing is significant. Every company that uses AI to automate or assist labor-intensive work will eventually face questions about that impact. The governance challenge is being able to answer those questions.
What Amodei actually proposed
Amodei's essay was notable for its candor about what frontier AI could do to labor markets. He argued that AI-driven displacement could be "larger and last longer" than the disruptions from previous technological transitions, and that governments should respond proactively rather than after the fact.
His specific proposals:
AI company taxation. Universal basic income could be funded through taxes on "relevant companies", AI developers and large deployers, or through increases to the capital gains tax. Amodei did not propose specific rates or a specific mechanism, but he explicitly identified AI companies as a logical source of funding given their role in driving displacement.
Retention tax incentives. Amodei called for incentives that would encourage employers to retain workers rather than replace them with AI. These would function similarly to the employee retention credits deployed during COVID, tax breaks for companies that maintain headcount despite having the capability to reduce it.
Wage insurance. Workers forced to take lower-paying jobs because AI reduced demand for their skills would receive wage insurance, partial compensation for the income gap. This is a concept that has been discussed in labor policy circles for decades but has rarely been implemented at scale.
Anthropic's $200 million research commitment is separate from these policy proposals but contextually related. The research program is designed to generate actual data on AI's labor market effects, sector by sector, skill category by skill category, which would provide the empirical foundation for any future policy intervention.
The Sanders bill
Senator Sanders' sovereign wealth fund legislation, introduced June 18, takes a different mechanism to a similar destination. Rather than a targeted AI company tax with proceeds directed to workers, Sanders' bill would create a sovereign wealth fund, a government-managed investment vehicle, financed by AI company taxes. The fund would distribute payments to displaced workers.
Sovereign wealth funds are common in countries with significant natural resource wealth (Norway, Saudi Arabia, Alaska). Sanders' proposal applies the model to AI productivity gains rather than resource extraction, on the premise that AI productivity gains are in part a public good, built on publicly funded research, internet infrastructure, and collectively generated data, that corporations are capturing privately.
The bill is in its early stages and faces significant legislative headwinds. But its introduction represents the first time a prominent U.S. senator has introduced a standalone bill to tax AI firms specifically for workforce displacement purposes, not just as general revenue.
The international dimension: what other major economies are doing
The Amodei proposal and Sanders bill did not emerge in a policy vacuum. Several major economies are already debating how to fund the social costs of AI-driven displacement, and the approaches vary significantly.
European Union. The EU AI Act does not address AI taxation directly, but EU labor ministers have discussed extending the EU Robot Tax proposal, first introduced by MEP Mady Delvaux in 2017, to cover AI systems that reduce labor demand at scale. The European Parliament's 2026 AI liability directive work has also raised the question of whether AI developers should contribute to a European Displacement Fund. No binding proposal has passed, but the debate is further advanced institutionally than in the US.
United Kingdom. The UK has taken a pro-innovation, light-touch approach to AI regulation under its post-Brexit framework. The government has explicitly rejected sector-specific AI taxes, arguing that taxing AI productivity gains would undermine the UK's competitiveness. However, the UK's Productivity Council published a 2025 report recommending expanded wage insurance for workers in sectors with high AI automation exposure, closer to the insurance-side of Amodei's proposal than the tax-side.
Canada. The Canadian government's 2025 AI and Society Strategy includes a commitment to study the fiscal impacts of AI on labor tax revenue, recognizing that if AI-driven productivity gains reduce payroll tax contributions (because fewer workers are employed), the funding base for social programs erodes. This framing, focusing on what AI displacement does to existing tax revenues rather than taxing AI companies directly, represents a third policy approach distinct from both the Sanders and Amodei models.
The divergence matters for multinational enterprises. A company operating across the US, EU, and UK will face different policy environments, and compliance programs built around one model may not translate. The most durable governance approach is tracking workforce impact data that can answer questions in any of these frameworks, regardless of which proposal eventually passes where.
The regulatory signal for enterprise governance
Neither proposal creates current compliance obligations. But both signal a direction that enterprise governance teams should anticipate:
AI workforce impact will become a disclosure category. The SEC has already moved toward requiring companies to disclose material AI risks. If AI-driven workforce displacement becomes politically salient, and the Amodei proposal suggests it already is, SEC guidance or rulemaking could extend to workforce impact disclosure. Companies that have never tracked which AI tools reduce headcount requirements, or by how much, will struggle to respond to disclosure requests.
State laws are ahead of the federal proposals. Minnesota's pending HF 4369 would require 90-day notice to workers before AI-driven displacement occurs. Connecticut and Illinois have active proposals on AI employment decisions. The federal proposals from Amodei and Sanders are framing documents; the actual near-term compliance risk is at the state level.
WARN Act obligations don't change, but AI doesn't insulate you. The federal WARN Act requires 60 days' notice before qualifying mass layoffs (50+ employees affected). AI automation as the cause of those layoffs does not exempt employers from WARN Act obligations. If your organization is planning headcount reductions enabled by AI automation, WARN Act counsel review is required the same as any other mass layoff.
The "we're not replacing workers, we're augmenting them" narrative is under scrutiny. Several large employers have used this framing while reducing headcount in roles adjacent to AI deployments. The Amodei proposal, the Sanders bill, and the broader policy discourse are making this framing harder to sustain without data. Enterprise governance teams that can point to concrete upskilling investments, retraining programs, and headcount impact analysis are better positioned than those that cannot.
What enterprise governance teams should do now
Inventory AI automation touchpoints. Identify which AI tools your organization uses that reduce the number of person-hours required for specific tasks. This is not the same as identifying which AI tools you use, most productivity AI doesn't directly reduce headcount. Focus on AI tools used in workflows where the alternative was dedicated human labor: document review, customer service triage, data entry, code generation for work previously handled by external contractors.
Establish a baseline. For each identified touchpoint, document the pre-AI headcount equivalent or labor hours. This creates a baseline for tracking and a foundation for any future impact reporting. Without a baseline, you cannot demonstrate what effect AI has had on your workforce, positive or negative.
Review retraining and upskilling investments. If your organization has invested in training workers whose roles are adjacent to AI automation, whether through internal programs or tuition support, document those investments. They are likely to become relevant to future policy discussions about employer responsibility for workforce transitions.
Monitor state legislation. Minnesota's 90-day notice bill and similar proposals in other states create near-term compliance risk, not future risk. The AI regulation deadline calendar tracks effective dates across major US AI laws and should be checked quarterly.
What enterprise teams should watch through the rest of 2026
The legislative environment around AI and workforce displacement will move faster in the second half of 2026 than it did in the first. Three specific things to track:
Economic Futures Research Fund outputs. The $200 million Anthropic-funded research program will publish findings that directly shape legislative debate. When initial data appears on AI's measured impact on employment and wages, it will become a resource that both sides of the legislative debate will cite. Teams with AI governance responsibilities should read it rather than relying on media summaries.
State action in MN, CA, and NY. State-level AI workforce legislation tends to move faster than federal action. California, Minnesota, and New York have the most active legislative calendars on AI employment issues. A bill that advances in any of them signals national momentum.
Sanders bill hearing schedule. Senate committee hearings on the sovereign wealth fund bill, if they happen, will generate testimony from AI companies, labor economists, and worker advocates that clarifies what specific policy mechanisms are actually being negotiated. Hearing testimony is often more revealing than the bill text about what legislative bargaining is actually occurring.
For HR teams managing both the current compliance obligations and the emerging regulatory horizon, the AI workforce displacement and WARN Act guide covers the obligations that are enforceable today while this federal debate continues.
For a governance framework covering the full range of AI policy developments, the AI governance checklist provides a structured starting point. The Dario Amodei broader policy analysis covers his earlier proposals on binding AI regulation.
Related Reading
- Dario Amodei AI regulation binding rules June 2026
- Great American AI Act: what the Obernolte-Trahan draft means for your business
- Pentagon Grok AI and human oversight: what Gillibrand's bill means for your policy
- One Big Beautiful Bill AI preemption: Senate voted 99-1 against it
- AI governance roles and responsibilities for small teams
- AI governance checklist 2026
- AI workforce displacement and the WARN Act: what HR teams need to know in 2026
