AI Now North Star Toolkit vs Big Tech's Data Center Heist

Every time a tech company announces a new data center, local residents don't get jobs — they get higher electricity bills. That's the central charge from the AI Now Institute (an independent research organization studying the social impact of AI), which just published two explosive documents: an op-ed labeling Big Tech's government lobbying a "bold, government-backed heist" and a practical policy toolkit for communities ready to push back.

The op-ed, "The Great AI Grift," dropped April 10, 2026. The North Star Data Center Policy Toolkit — a step-by-step playbook for local and state governments — came out April 1, 2026. Together, they represent the most organized counterattack yet to the AI infrastructure boom narrative sweeping Washington and Silicon Valley.

Arms-Race Logic and the "National Security" Excuse

The core of AI Now's argument: tech companies aren't asking for investment — they're demanding unconditional government support by weaponizing the China threat. By comparing AI expansion to the Apollo program and the Manhattan Project, the industry recasts any attempt to regulate or constrain corporate power as unpatriotic and anti-competitive.

AI Now's verdict: "The reality looks more like a bold, government-backed heist." The goal, the Institute argues, is to extract maximum public subsidies — tax breaks, infrastructure, regulatory rollbacks — while shifting the actual costs onto people who live near these facilities: environmental damage, resource depletion, higher utility bills, community disruption.

The pattern AI Now documents in detail:

• Tax incentives that drain municipal budgets while promised local hiring consistently fails to materialize
• Fossil fuel increases as data centers demand always-on power regardless of local clean energy commitments
• Electricity rate hikes for ordinary residents as AI workloads spike grid demand
• Water depletion in drought-prone regions, as cooling systems (the industrial fans and liquid systems that prevent servers from overheating) consume millions of gallons annually
• Regulatory capture — the process by which corporations gain influence over the agencies meant to oversee them — advancing under the banner of "national competitiveness"

The North Star Toolkit: A Practical Resistance Playbook

Published April 1, 2026, the North Star Data Center Policy Toolkit is deliberately practical. It doesn't just describe the problem — it gives policymakers and organizers specific tools, each rated by implementation strength: "very strong," "strong," or "weak," based on legal feasibility in different jurisdictions.

The toolkit contains four main components:

• Local policy recommendations for stopping or slowing new data center permits, with feasibility ratings per measure
• State-level interventions covering energy regulation, water rights, tax incentives, and air quality standards
• Issue-specific modules for communities with particular vulnerabilities — water scarcity, grid instability, air quality — rather than one-size-fits-all rules
• "North Star" proactive standards for jurisdictions not yet hosting large data centers, establishing protections before the first permit application arrives

The Institute is candid about limits: "Not every intervention will be feasible in every locality, state, or region due to differences in local and state laws, existing regulation, and political conditions." But the toolkit is built for coalitions — groups of advocates, community organizations, and elected officials working together — not single officials acting alone.

Communities in active conflicts with data center developers can reach the Institute directly at datacenters@ainowinstitute.org.

How "Impact Lingo" Hides the Real Costs

A companion essay — "Beyond Impact Lingo: Questioning, Concretizing, Building" (March 12, 2026) — documents something more subtle: the systematic co-option of socially meaningful language to market extractive practices. The research draws on 12 interviews with leading critical AI researchers, including:

• Karen Hao — journalist covering AI's social implications at The Atlantic
• Timnit Gebru — AI ethics researcher and founder of the Distributed AI Research Institute (DAIR)
• Audrey Tang — former Taiwan Digital Minister and civic technology leader
• Meredith Whittaker — president of Signal and longtime tech accountability advocate

Their collective finding: terms once carrying real political meaning have been hollowed out and repurposed as marketing instruments:

• "AI for Good" — deployed as brand positioning without measurable targets, timelines, or accountability mechanisms
• "Democratization" — originally meant redistributing power; now used to mean "making products available for purchase"
• "Sovereignty" — stripped from self-determination movements, rebranded to mean US-controlled national AI infrastructure
• "Open source" — invoked to signal community values while actual code access remains restricted or commercially licensed
• "Frugal AI" — efficiency framing used to justify cheaper, less regulated models in Global South markets

The end result, the Institute warns, is what it calls "a 'machine god' future that thrives on environmental exploitation, mass exclusion, and modern imperialism" — a deliberately stark description of the current trajectory if left unchallenged. The Institute also calls out "vibes-based policymaking" (decisions made on intuition and industry pressure rather than empirical evidence), and demands evidence-grounded policy across the board.

2,000 Languages Left Behind

One of the report's sharpest statistics: more than 2,000 African languages need dedicated linguistic AI systems (software trained to understand, translate, and generate text in a specific language) — and are being largely ignored by AI companies focused on English-language markets where commercial returns are highest.

This isn't a gap that "AI for All" rhetoric is closing. The concentration of compute, investment, and research talent in high-income markets means that communities generating the least profit continue to receive the least benefit — even as they increasingly bear the environmental and economic costs of data center expansion in their regions. The Institute frames this as structural, not accidental: a system optimized for investor returns will consistently under-serve populations that can't generate large returns, regardless of how many "AI for Africa" press releases get published. Follow our AI policy and data center news to track these global developments.

Resistance Is Already Working — Where to Start

The toolkit highlights documented wins in Chile, multiple US states, and Canada, where organized local resistance has achieved measurable policy outcomes. The common pattern: communities that mobilized early — before construction broke ground — achieved the strongest protections. Those that waited until operations began had far fewer tools available.

Three entry points depending on where your community stands today:

1. No data center planned yet? Use the toolkit's "North Star" proactive standards now. Establishing baseline protections before any company files a permit application is the strongest legal and political position to be in.
2. Data center approved or under construction? Focus on the local policy section. Start with measures rated "strong" or "very strong" — weaker measures can still create accountability mechanisms and build coalition momentum even if they can't stop construction outright.
3. Already in an active fight? Contact the Institute directly at datacenters@ainowinstitute.org — they're documenting active cases and connecting communities facing similar pressures.

AI Now's overarching message is direct: AI infrastructure expansion is not inevitable, and it is not uniformly beneficial. Communities that have accepted it as such have consistently gotten the worst of both — the pollution and the broken promises. The North Star Toolkit exists to change that equation, one city council meeting at a time. You can access the toolkit here and watch for local permit filings in your area before the window for proactive policy closes.

Related Content — Set up AI automation tools | AI automation guides | More AI policy news