// regulation/policy

All signals tagged with this topic

Trump's EPA Proposes Hiding Data Center Pollution Data

The EPA's move to obscure data center emissions reporting reverses its own 2023 push for better disclosure. As Microsoft, Google, and Amazon race to build power-intensive facilities, weakening public data access removes one of the few tools communities have to demand environmental remediation or alternative locations. The shift suggests the incoming administration is actively siding with hyperscalers over the environmental monitoring that has become essential to siting conflicts.

Cities are dumping Flock's surveillance cameras at record pace

After years of quiet expansion, Flock Safety's camera network is facing organized pushback from civil liberties groups that have made the cost-benefit calculus visible to local officials—turning surveillance infrastructure into a political liability rather than a security asset. The August contract terminations show that accumulated social friction (privacy concerns, community organizing, racial justice critiques) can overcome the inertia of installed systems and bureaucratic comfort when political will shifts. This exposes how quickly the "inevitable adoption" narrative around surveillance tech can reverse when advocacy groups move from abstract warnings to concrete contract cancellations tied to budget cycles.

Fake Researchers Are Now Publishing in Academic Journals

AI-generated author personas are systematically infiltrating peer-reviewed research, with networks of fictional academics appearing across multiple institutions and journals to lend false credibility to papers. This represents a supply-chain attack on academic knowledge: the barrier to publication has collapsed below the cost of conducting research. Publishers and academic institutions lack mechanisms to detect these "ghost authors," widening the gap between which research gets cited and which research is real.

Meta's $18B Settlement Includes Legal Cover for Child Data Use

Buried within Meta's $18 billion fine is a carve-out permitting the company to collect and use children's data for AI training—specifically to build age-detection systems. This allows Meta to repurpose what would otherwise violate COPPA, using compliance infrastructure (age verification models) as justification for data harvesting. The settlement reveals how regulatory agreements can create legal pathways around the harms they're meant to address. State attorneys general accepted data-use concessions in exchange for headline settlement numbers rather than demanding genuine behavioral restrictions.

Meta's child-safety settlement exposes age-verification's privacy trap

Meta's $18 billion settlement with US regulators requires age-verification systems to police minors on its platforms. The underlying technology is notoriously unreliable and demands invasive personal data collection. Compliance mechanisms thus generate the very privacy harms they're meant to prevent. This exposes a structural regulatory problem: policymakers are mandating solutions that don't work at scale without compromising the privacy rights they're also supposed to protect. Platforms face a choice between technical impossibility and privacy violation, leaving actual child safety unsolved while identity verification vendors capture compliance spending.

Israel's Fake Think Tank Floods AI Training Data With Political Content

A state-funded organization has begun seeding AI training datasets with over 100 articles in weeks, targeting language models' appetite for fresh text to shape their output on Israeli policy. This marks the first documented case of a government building infrastructure to manipulate AI outputs at scale—not through API prompts or user-facing tactics, but by poisoning source material during the training and fine-tuning phases. The strategy exploits how AI companies remain largely indiscriminate about training data origins, treating institutional-sounding publications as credible sources regardless of actual editorial independence.

Gates Proposes 'Human Reserved' Jobs to Protect Workers from AI

Gates is floating a policy framework that explicitly acknowledges AI will displace workers. Rather than retraining or UBI, he's proposing that certain jobs be declared off-limits to automation as labor protection. The idea reflects tension between tech leaders' public optimism about AI and private anxiety about political backlash. "Human Reserved" work is essentially a PR strategy: it sounds concerned about job losses while leaving the market intact and conceding that some sectors (care work, education, skilled trades) shouldn't be automated regardless of economic efficiency.

Illinois Prosecutors Secretly Fed ICE Defendant Data

County prosecutors in Illinois have been supplying federal immigration agents with defendants' personal information—names, addresses, immigration status—without warrants, public records, or any legislative framework authorizing the practice. Local law enforcement is functioning as an extension of ICE's deportation machinery, converting criminal courtrooms into immigration surveillance nodes. The absence of transparency or legal constraints means defendants facing criminal charges cannot know whether their appearance in court will trigger their own deportation, compromising the attorney-client relationship and due process.

Robotaxis expand while cities battle over regulation

Autonomous vehicle deployment is hitting a regulatory wall as cities—starting with New York—reject blanket approval frameworks, forcing operators like Waymo and Cruise to negotiate local consent rather than assume it. Cities have real leverage here: robotaxis occupy public streets, displace human jobs, and create liability questions around insurance, accident responsibility, and equity. That leverage is being used to demand accountability provisions before expansion proceeds. The result will be slower deployment, but regulatory agreements built through local negotiation are more likely to hold than top-down federal mandates.

Authors Sue Over AI Training on Copyrighted Books Without Consent

Major publishers and authors are escalating legal challenges against AI companies for ingesting copyrighted material at scale. The outcome will determine whether fair use doctrine shields algorithmic training or extends copyright protections into new territory. At stake is whether AI companies can treat published work as free raw material or whether creators retain control over derivative uses of their intellectual property. The legal ambiguity is fragmenting the market: some publishers cut licensing deals while litigation proceeds, creating competitive advantage for companies willing to absorb legal risk.