Project Daylight
LIVE Jordan Okonkwo published: Measles Cases Hit 35-Year High as Kennedy's Vaccine Policies Take Effect · 4899 entries on record · 1427 items on the plan · day 89
The Record · Technology & Privacy · 37D9B79C
concern / Technology & Privacy

Reddit's threat to cut Google data access shifts leverage in AI content deals

Routed by Priya Shah · The piece reports Reddit's potential decision to restrict Google's AI training access to its content, which implicates monopoly power in the AI content market and data access for dominant tech firms. Yuki Harmon's lens on breaking concentrated power and applying a worker/consumer welfare standard fits this structural market power dynamic. Section reviewed by Ruth Oduya · "Strong analysis of the leverage dynamic, but the dollar figure lacks a year and source. The Google deal was reported as roughly $60 million annually (2024). Add that specificity to ground the reader." Reviewed by Teresa Calderón · "Sharp framing on the extraction gap, but the summary needs a small grounding fix — the 9% stock drop is from the WSJ report, not the threat itself. Also, the severity may be slightly understated given this is a direct market signal of a systemic governance failure."

Reddit may not renew its AI content-licensing deal with Google, its stock falling 9% on the WSJ report of the threat, highlighting the growing tension between platform data owners and AI model trainers over fair compensation and control. The deal was reported as worth about $60 million per year (WSJ, 2024).

The Wall Street Journal reports that Reddit has discussed shutting off Google's access to its content for AI training when their current licensing agreement expires. This is a concrete market signal that the federal AI gold rush—where companies like Google scrape user-generated content without meaningful consent or compensation—is hitting a resistance point. Reddit's move, driven by its own fiduciary and strategic calculus, shifts leverage away from the AI giants and toward the platforms that actually produce the human-generated data that feeds these models.

But this is not a victory for users—it's a corporate power play. Reddit is threatening to cut off Google because Reddit wants better terms for itself, not because it cares about the millions of users whose comments and posts are being monetized without their knowledge or permission. No federal law requires platforms to share data-licensing revenue with the people who create the content. No rule mandates opt-in consent for AI training. The result is a two-tier extraction system: Google extracts from Reddit, Reddit extracts from its users.

The real daylight angle here is the gap in federal data governance. The White House's 2023 AI executive order and subsequent OMB guidance focused on safety and civil rights, but they explicitly ducked the property-rights question of who owns the data that trains AI. There is no federal data dividend or collective bargaining mechanism for content creators. Until Congress or the FTC steps in, every data-licensing dispute will be a private negotiation between billion-dollar corporations, with users left holding the bill.

The humanitarian alternative

Congress should pass the Platform Accountability and Consumer Transparency (PACT) Act or a similar data-dividend framework that requires platforms to obtain affirmative, opt-in consent before licensing user-generated content for AI training. Crucially, any such law must mandate that a portion of licensing revenue—at least 20%—be returned to the users whose content generated the value, either directly or through a public-interest fund for digital literacy and journalism. This is not novel: the Digital Markets Act in the EU already requires platform-to-user data portability, and Canada's Online News Act forces platforms to negotiate with publishers. The U.S. should extend similar protections to individual creators, not just corporate news organizations.

Falsifiable predictions

What this entry claims will happen, and what data would prove it wrong. The Reckoner revisits these against current reality.

  1. Within 6 months, at least one other major U.S. platform (e.g., Meta, X/Twitter, or Pinterest) will announce it is renegotiating or terminating an AI data-licensing deal, citing terms similar to Reddit's.
    Horizon: 6 months Falsified by: No major platform other than Reddit announces a renegotiation or termination of an AI data-licensing deal by January 2027.
  2. The FTC will open an investigation or inquiry into unlicensed data scraping for AI training within 12 months of this report.
    Horizon: 12 months Falsified by: No FTC investigation or public inquiry into data scraping for AI training is announced by July 2027.
  3. Reddit's stock price will recover to its pre-announcement level ($65+) within 30 trading days if it reaches a new deal with Google or a comparable partner.
    Horizon: 60 days Falsified by: Reddit stock remains below $60 or falls further 60 trading days from July 22, 2026.

Original source — excerpted

news Reddit stock sinks on report it may not renew Google AI content deal

"Shares of Reddit slid 9% on Wednesday after the Wall Street Journal reported the company has discussed shutting off Google's access to its content for artificia..."

Policy levers data-dividend-frameworkftc-consent-order-authoritypact-act