📢 Data center messaging | AI+ Government Summary

1. Key Themes

Theme 1: Democrats Are Searching for a "Business-Friendly" AI Infrastructure Position

State-level Democrats are actively workshopping a political lane on data centers that avoids being explicitly anti-business while still addressing constituent concerns around affordability, energy, and workforce disruption. New York is the test case.

"The 'thoughtful' approach for Democrats is to avoid demonizing business and just focus on protecting their constituents." — Gov. Kathy Hochul

"New York is emerging as a test bed for one potential approach."

Theme 2: AI Training Data Is Becoming a Competitive Moat and Regulatory Target

The fight over AI training data is evolving from a copyright dispute into an antitrust and competition issue. Physical destruction of books by dominant AI companies is now being framed as an anti-competitive act that forecloses rivals from accessing necessary inputs.

"The fight over AI training data may move from copyright into competition. That could include regulators scrutinizing whether dominant AI firms are literally eliminating resources their rivals need to compete."

Civil society groups argue that AI companies buying, scanning, and destroying books constitute "a destructive new data acquisition practice by dominant AI companies."

Theme 3: State Governments Are Filling the Federal Consumer Protection Vacuum

With the CFPB gutted under the Trump administration, state attorneys general are stepping into the breach on AI-related consumer protection — including employment, fair competition, and disclosure requirements.

"The Consumer Financial Protection Bureau has been gutted during President Trump's second term. It has lost much of its workforce and the agency has dropped dozens of investigations and transparency practices, prompting Democratic and Republican attorneys general to take on more consumer protection work."

"States have also been moving forward on AI regulation, while Washington has focused on a deregulatory approach coupled with preempting state action."

Theme 4: AI Workforce Anxiety Is a Real Political and Operational Variable

Frontline workers — not just executives — are articulating nuanced, on-the-ground concerns about AI displacement. The emotional labor and irreplaceable human judgment in certain jobs is being surfaced as a policy argument.

"You need human interaction in some things. You may get a phone call that someone just got the worst news of their life and they're crying and I'm trying to help them understand this cancer diagnosis." — Lynn Brooks, claims processor and union steward

"AI is a tool. I see the need for it. But I don't want it to replace the family." — Lynn Brooks

2. Contrarian Perspectives

Perspective 1: Book Destruction Should Be Viewed as an Antitrust Issue, Not Just a Copyright One

The conventional framing of AI companies using copyrighted books for training is a copyright/IP dispute. Civil society groups are now arguing this is actually a competition issue — incumbents are physically destroying a shared resource to starve challengers of training data, effectively building a moat that regulation hasn't yet addressed.

"We urge the FTC to view this practice not in isolation but rather as the latest escalation in a documented pattern of anticompetitive conduct designed to create an insurmountable systemic moat around AI incumbents."

The Washington Post reported, citing court filings, that "Anthropic spent millions to acquire and physically remove the spines of books to feed their scanned pages into Claude."

Perspective 2: A Short Moratorium May Be the Most Business-Friendly Move Available to Democrats

Counter-intuitively, Hochul's one-year moratorium — while seemingly restrictive — positions her to the right of most progressive Democrats and preserves her business relationships. The moratorium signals control without triggering the long-term contingency-based bans other progressives favor.

"Hochul will forever be able to boast that she was the first governor in the nation to sign a data center moratorium, but there is still a meaningful distance between her and progressives that allows her to maintain her business relationships."

Progressives want to "make the construction contingent on achieving AI safety and other much loftier goals" — whereas Hochul's is a bounded, one-year pause.

Perspective 3: Government Telling Tech Companies How to Market Data Centers Is Overreach — Even for a Skeptical Democrat

Hochul pushed back on the Senate GOP's effort to pressure AI companies on their public messaging around data centers, suggesting there's bipartisan overreach happening on both sides of the AI debate.

"I don't think it's government's role, certainly not the United States Senate's role, to be telling companies how to market their products."

3. Companies Identified

Anthropic

Description: Major AI lab, maker of Claude
Why mentioned: Cited in Washington Post reporting (via court filings) as having physically destroyed books to scan and train its models
Quote: "Anthropic spent millions to acquire and physically remove the spines of books to feed their scanned pages into Claude."

Google, Microsoft, OpenAI

Description: Dominant AI incumbents
Why mentioned: Cited alongside Anthropic as having "faced similar copyright lawsuits" related to AI training data practices
Quote: "Google, Microsoft and OpenAI have faced similar copyright lawsuits, per the Post."

Highmark Blue Cross Blue Shield of Western New York

Description: Regional health insurance company
Why mentioned: Employer of Lynn Brooks, whose testimony illustrated real-world AI workforce anxiety in emotionally sensitive jobs
Quote: Used as a case study in Hochul's workforce development roundtable

Teleport (Sponsor)

Description: Cybersecurity infrastructure company
Why mentioned: Sponsor; promoting "Agent Trust" framework for securing AI agents beyond traditional Zero Trust models
Quote: "AI agents are neither human nor machine — they act at machine speed with human unpredictability, and can produce globally destructive outcomes."

4. People Identified

Kathy Hochul

Description: Governor of New York (Democrat)
Why mentioned: Pioneer of the "business-friendly" Democratic data center moratorium approach; being watched as a model for other Democrats nationally
Quote: "I want Democrats to win, and if this [approach] helps them, embrace it, right?"

Nathan Johnson

Description: Democratic nominee for Texas Attorney General
Why mentioned: Proposing a 30-day "AI audit" within his first days in office to recommend legislative updates to consumer protection, fair competition, and employment laws
Quote: "Stuff that we thought the statutes were fine to handle three years ago might as well have been 3,000 years ago. The pace is changing so rapidly."

Lynn Brooks

Description: Claims processor for Highmark Blue Cross Blue Shield of Western New York; union steward
Why mentioned: Frontline worker whose testimony at Hochul's roundtable articulated the irreplaceable human dimension of emotionally sensitive jobs
Quote: "You need human interaction in some things... AI is a tool. I see the need for it. But I don't want it to replace the family."

Ken Paxton

Description: Current Texas Attorney General (Republican)
Why mentioned: Referenced as the incumbent whose seat Nathan Johnson is running to capture
Quote: N/A — mentioned by name only as context

5. Operating Insights

Insight 1: Employers Who Use AI to Make Employment Decisions Face an Emerging Disclosure Gap

Even in states that have passed AI-specific laws, a meaningful compliance gap persists. Texas, for example, doesn't currently require employers to disclose AI use in termination decisions — and Johnson's proposed audit would target exactly this type of gap.

"Texas has passed AI-specific laws, but Johnson says they still fall short — for example, by not requiring employers to disclose when they use AI to determine who gets fired."

Takeaway for operators: Proactively audit and document AI's role in HR decisions now, before disclosure mandates arrive at the state level.

Insight 2: Political Positioning on AI Infrastructure Is an Active Opportunity — With a Narrow Window

The competitive space for a coherent, business-compatible AI policy position is still being defined. Governors, AGs, and legislators who establish frameworks early gain durable credibility — and first-mover advantages like Hochul's.

"As candidates across the country scramble for a winning position on the issue, New York is emerging as a test bed for one potential approach."

Takeaway for operators/lobbyists: Engage state-level policymakers now, before positions harden. The window to shape "thoughtful" frameworks is open.

6. Overlooked Insights

Insight 1: The FTC's Jurisdiction Over Training Data Practices Has Not Been Tested

The civil society groups are asking the FTC to classify physical book destruction as an "unfair method of competition" — an untested application of FTC authority that, if pursued, would represent a major expansion of how antitrust law is applied to AI inputs. This is not a copyright claim; it's a competition claim, and that distinction matters.

"The groups want the FTC to determine whether such conduct constitutes an unfair method of competition, arguing that any AI company that does so is 'starving the market' of critical source materials."

Insight 2: The Senate GOP Is Actively Pressuring AI Companies on Data Center Marketing

Briefly noted but significant: the Republican Senate campaign arm issued a memo pressuring AI companies to improve the public optics of data centers — a rare instance of the GOP using political leverage over the industry it nominally champions. This signals that data center backlash is becoming bipartisan at the electoral level.

Hochul "blasted the Senate GOP campaign arm's pressure on AI companies to fix the optics of data centers," calling it inappropriate government overreach into how companies market their products.