Communities force AI governance to the physical layer

Infrastructure as regulation. Voters in 13 states rejected data-center projects in roughly 40 to 50 local ballot measures this November, asserting control over the physical foundations of AI expansion (AI Governance, Ethics and Leadership). Dina Powell McCormick of Meta noted that this democratic backlash makes new U.S. build-outs significantly more challenging, as communities prioritize concerns over local energy costs and strained power grids (Raw Egg Nationalist from In the Raw).

Legislative pressure mounts. Democrats in Congress plan to introduce bills requiring independent testing of advanced AI systems and establishing corporate liability for AI-driven harm. These moves follow a year where U.S. employers attributed approximately 120,000 job cuts to AI, fueling political momentum for expanded unemployment support and mandatory workforce-impact disclosures (AI Governance, Ethics and Leadership).

The governance gap. Most AI agents currently operate without identity, ownership, or a kill switch, leaving 95% of organizations unable to contain an agent once it is compromised (Cashless). While some firms move to cut costs by shifting to open-weight models, the lack of a standardized liability framework means users remain legally exposed when autonomous systems act in ways that cause real-world damage.

• Anthropic’s use of AI for R&D has accelerated its capability growth on well-specified tasks to roughly twice the rate seen in 2024 and 2025 (ControlAI).
• Global corporate spending on AI is projected to reach $2.7 trillion in 2026, a 49.5% increase that strains energy capacity for other sectors (MMT101.ORG).