Global Tech Policy Tracker — 2026-06-15

The U.S. House of Representatives unveiled a bipartisan AI regulation bill that would preempt state laws for three years—a major move to federalize AI governance that has drawn both industry praise and consumer advocate criticism. Meanwhile, the EU continues refining its AI Act implementation with deadline extensions for high-risk systems, and the UK Parliament published fresh guidance on AI regulation frameworks. -->

Top Story

2 sources

whitehouse.gov

whitehouse.gov

whitehouse.gov

whitehouse.gov

House Bipartisan Draft Proposes Federal AI Framework, Preempting State Laws

On June 4, 2026, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a 269-page discussion draft of the "Great American Artificial Intelligence Act of 2026"—the first comprehensive federal AI bill to reach Congress. The draft proposes a three-year preemption period during which states would be prohibited from enforcing their own AI model development regulations, a move intended to prevent a fragmented regulatory landscape and preserve federal oversight.

The bill covers frontier model safety, workforce protections, and national security vetting of advanced AI systems. Tech companies including OpenAI and others have largely supported federal preemption to avoid compliance with dozens of state-level rules. However, labor unions and consumer rights groups have raised concerns that federal preemption could weaken worker protections and limit states' ability to regulate high-risk AI applications in their jurisdictions. This draft represents the most significant legislative effort yet to establish a unified U.S. AI policy before the 2026 midterm elections.

1 sources

politico.com

politico.com

New Legislation & Regulatory Actions

United States: Great American Artificial Intelligence Act of 2026 (Discussion Draft)

• What happened: House members Obernolte and Trahan released a bipartisan 269-page discussion draft proposing comprehensive federal AI regulation, including provisions on frontier model safety, vetting of advanced AI systems, and workforce impacts.
• Who it affects: AI developers, frontier model creators (OpenAI, Anthropic, Meta, etc.), tech workers, and consumer groups seeking AI safeguards.
• Status: Discussion draft released June 4, 2026; not yet formally introduced as legislation. Expected to advance before midterm elections.
• Why it matters: First major attempt at federal AI preemption. A three-year preemption clause would block states from regulating AI model development, creating uniform national rules instead of a patchwork of state laws.

EU: Cloud and AI Development Act (Proposed Regulation)

• What happened: The European Commission and EU Parliament are advancing a new "Cloud and AI Development Act" (Regulation 2026/XXXX) to establish cloud sovereignty rules and boost AI infrastructure investment across member states.
• Who it affects: Cloud providers, AI infrastructure companies, and member states seeking digital sovereignty and reduced dependence on non-EU computing resources.
• Status: Proposed; expected to complement existing AI Act enforcement timelines.
• Why it matters: Signals EU focus on building competitive AI infrastructure while implementing the AI Act. The regulation strengthens Europe's digital autonomy in critical AI development capabilities.

UK: AI Regulation Framework Update

• What happened: The UK House of Commons Library published updated guidance on AI regulation in the UK, clarifying the sectoral approach to AI governance across financial services, healthcare, and other regulated industries.
• Who it affects: UK-based AI developers, financial services firms, healthcare organizations, and regulated sectors relying on AI deployment.
• Status: Published June 15, 2026 (latest available). Non-binding guidance; part of ongoing UK framework development.
• Why it matters: The UK continues its "light-touch, sector-specific" approach contrasting with the EU's prescriptive AI Act. Provides clarity to industry on which regulators have AI oversight in their sectors.

Enforcement & Penalties

• EU AI Act Compliance Fines: Under Article 99 of the EU AI Act, companies face administrative fines up to €35 million or 7% of global annual revenue (whichever is higher) for prohibited AI practices, and 3% of global revenue for high-risk non-compliance. These penalties exceed typical GDPR fines and are designed to demand board-level attention to AI governance.

• GPAI Systemic Risk Requirements: General-Purpose AI (GPAI) models exceeding 10²⁵ FLOPs during training face mandatory adversarial testing, incident reporting, and enhanced cybersecurity requirements under the EU Act. Non-compliance carries escalating penalties.

Industry Response

• OpenAI Diverges from White House AI Safety Approach: On June 3, 2026, OpenAI published its own regulatory framework for advanced AI that differs from the Trump administration's voluntary vetting and enhanced intelligence community involvement. OpenAI proposes alternative governance structures, signaling industry pushback on certain federal policy directions.

• Tech Industry Supports Federal Preemption: Major tech companies have praised the Obernolte-Trahan draft bill's three-year preemption clause as a necessary step to prevent regulatory fragmentation across 50+ state AI laws and enable national innovation. However, no major company has formally testified before Congress on the draft.

• Labor & Consumer Groups Oppose Preemption: Labor unions and civil society organizations including the ACLU have publicly criticized the preemption language, arguing it would prevent states from enacting worker protections and high-risk AI safeguards. These groups advocate for complementary state and federal rules rather than federal override.

Region Scorecard

Analysis: What This Means

• Federal vs. State Tension Peaks: The Obernolte-Trahan preemption clause would eliminate a three-year window of state innovation in AI regulation. Companies currently navigating 50+ state laws will face immediate pressure to comply with federal rules instead—but only if Congress passes the bill. For now, plan for continued state-level enforcement (California, Colorado, etc.) alongside federal uncertainty.

• EU Penalties Now Exceed GDPR: With fines up to 7% of global revenue for prohibited AI uses, the EU AI Act's enforcement teeth are sharper than GDPR's. Any company deploying prohibited AI systems (e.g., real-time biometric identification) in EU markets faces existential financial risk. Audit your AI systems against Article 5 prohibitions immediately.

• Divergence Between U.S. and EU Governance Models: The U.S. draft favors a unified, preemptive federal approach; the EU is building layered Cloud and AI Development rules alongside the AI Act. Companies must maintain dual compliance strategies—lighter federal requirements in the U.S. (if passed), stricter sectoral and AI Act compliance in the EU.

• Industry Fragmentation on Preemption: While tech firms support federal preemption to cut compliance costs, labor and civil society oppose it. This coalition split will likely slow congressional momentum. Expect lobbying intensity to increase over summer 2026.

What to Watch Next Week

• House Judiciary Committee response to the Obernolte-Trahan draft (expected mid-June).
• EU AI Act enforcement timeline confirmation following May 7, 2026 Digital Omnibus negotiations (Annex III high-risk systems deferred to December 2027; transparency requirements hold at August 2, 2026).
• OpenAI and other major AI labs' formal regulatory positions published as testimony or white papers ahead of potential congressional hearings.

Editor's Note: This report covers developments published after June 8, 2026. Earlier announcements from June 1–8 (including the White House executive order on AI security) are documented in prior issues. Regional activity reflects public announcements only; private enforcement actions and confidential regulatory guidance may not be captured.