Skip to content

Tuesday Brief · Friday, October 23, 2026

Google's Local News AI Deals, USA Today's $250M Lawsuitand What AI Really Sees Behind Your Paywall
The Fourth Estate AI Brief

Was this forwarded to you? Get The Fourth Estate AI Brief in your inbox →

Trevor Slette

Trevor Slette

Co-founder, Quadd.ai · Reply to me at trevors@quadd.ai

View this issue in your browser

Tuesday Brief

Google's Local News AI Deals, USA Today's $250M Lawsuitand What AI Really Sees Behind Your Paywall

The interesting thing this week with AI and publishing; the money is finally showing up, but the terms are getting complicated. Google expanded its News AI pilot to 400 publishers including local newsrooms in Charleston, Casper, and the Illinois-Iowa corridor. USA Today filed a $250 million lawsuit against OpenAI. And a piece making the rounds this week asks a question every gated-content publisher should be asking: if AI can't see your best work, does it exist?

GOOGLE'S LOCAL NEWS PLAY

Google expands AI pilot to 400 publishers, adds local U.S. newsrooms

The Rundown: Google's News AI pilot now includes The Post and Courier in Charleston, Oil City News in Wyoming, and Shaw Media's Illinois-Iowa chain — alongside CNN and the Financial Times.

The details:

  • Program grew from initial December launch partners to 400+ publishers globally
  • Local participants paid flat licensing fees, not pay-per-value based on how much content feeds AI outputs
  • Pay-per-value model applies separately to AI Overviews, AI Mode, and Gemini products
  • AI features include 'story previews' (AI-generated summaries) and 'audio briefing' (daily podcast in Google News app)
  • Google declined to disclose financial terms of any licensing agreements

Why it matters for us: If Google calls your 15,000-circ daily, this tells you the deal structure: flat fee upfront, no backend based on actual usage. Negotiate that number hard — once you sign, there's no meter running.

Read at news.lavx.hu →

USA TODAY VS. OPENAI

Gannett papers file $250 million lawsuit against OpenAI over ChatGPT training

The Rundown: USA Today and 19 affiliated newspapers are suing OpenAI for copying hundreds of thousands of articles without permission and asking the court to destroy the AI models that contain them.

The details:

  • Complaint filed October 8, 2026 in SDNY identifies 160,000+ entries from newspaper websites in OpenAI's training data
  • Approximately 83,300 entries came from USA Today's flagship site alone
  • Publishers tested GPT-5.6 and allege it reproduced protected reporting when asked to summarize specific articles
  • Lawsuit claims OpenAI collected articles regardless of paywalls and deliberately targeted news content
  • Case joins consolidated litigation including The New York Times lawsuit from December 2023

Why it matters for us: The 160,000-article inventory and $250M damages claim give smaller papers a concrete benchmark for calculating what their archives might be worth in licensing negotiations — or in court.

Read at lawcommentary.com →

THE PAYWALL PARADOX

If your best content is gated, AI search pretends it doesn't exist

The Rundown: A piece circulating this week makes a point worth sitting with: LLMs can't fill out forms or bypass paywalls, so gated content gets excluded from both AI training and real-time retrieval entirely.

The details:

  • AI crawlers only see splash pages and teasers, the actual content behind gates is invisible
  • Gated PDFs, white papers, and archive content get excluded from AI training datasets and retrieval systems
  • When AI delivers zero-click answers, invisible content means competitors get cited instead
  • Hybrid approach suggested: publish summaries and key data ungated, gate only the full download

Why it matters for us: If you gate your investigative pieces or archive content behind registration walls, AI tools recommending local news sources will cite your competitors instead. Worth auditing which content actually needs a gate versus which should be discoverable.

Read at marketingadvice.substack.com →

MONITORING YOUR AI REPUTATION

Tokyo media company launches tool to track how AI describes your brand

The Rundown: TNL Mediagene announced CiteRadar, a SaaS platform that monitors what AI models say about brands and products, addressing a visibility gap most publishers haven't thought about yet.

The details:

  • CiteRadar tracks brand mentions across major AI models and identifies narrative shifts
  • Platform provides automated intelligence reports on how AI describes companies
  • Beta expected in 2026 after proof-of-concept with enterprise customers
  • Company also launched Agentic Newsroom for automated translation and distribution across Asian markets

Why it matters for us: When someone asks ChatGPT 'what's the best news source in your county,' what does it say about your publication? That's a question worth asking and tracking, even if you build the monitoring yourself.

Read at tnlmediagene.com →

SOCIAL MEDIA NUANCE

Gallup finds moderate social media use linked to better wellbeing than quitting entirely

The Rundown: The 2026 State of the World's Emotional Health report surveyed 97 countries and found something counterintuitive: low social media users show worse life satisfaction than moderate users.

The details:

  • 81% of adults worldwide used social media in the past week; 95% among ages 15-29
  • Moderate use (30 min to 2 hours) associated with fewer negative emotions than high use
  • Low use correlated with lower life satisfaction than moderate use, not higher
  • High use (2+ hours) linked to more anger, loneliness, sadness, stress and worry across all ages
  • 47% of adults 15-29 are high users versus just 12% of those 50+

Why it matters for us: When covering school board debates about phone bans or local mental health initiatives, these numbers add nuance — 'ban it all' policies may not improve outcomes if low users actually report worse life satisfaction.

Read at news.gallup.com →

THE UPDATE

Quick hits from the week

The Rundown: Smaller items worth a glance.

The details:

  • OpenAI's fair-use defense in the USA Today case will be tested against 160,000 documented training entries — a scale that makes 'transformative use' arguments harder to sustain.
  • Google's audio briefing feature stitches headlines into daily podcasts with source citations — watch whether this drives traffic or replaces the need to click through.
  • The CiteRadar concept raises a question every publisher should be asking internally: what does Claude say about your paper when someone asks for local news recommendations?

Why it matters for us: These are the threads worth watching over the next few weeks.

That's the week. The licensing deals are getting more specific, the lawsuits are getting bigger, and the question of what AI can and can't see behind your paywall is worth ten minutes of your Wednesday morning. Trevor Slette runs Quadd.ai — AI tools built for publishers.

Trevor

Trevor Slette

Co-founder, Quadd.ai · 28-year community publisher

trevors@quadd.ai · Book a 15-min call · LinkedIn

Found this useful? Forward to a colleague →

From meeting audio to usable copy.

Upload a council meeting, an interview, a coach's post-game scrum. Get a clean transcript and an AI summary, then ask Quadd to compose a first-draft story or paste your own prompt, and get formatted output in your publication's exact style. No more scrubbing through forty minutes of audio at 9pm because the lede quote was somewhere in the middle. Trained on the way reporters actually use audio, not the way podcast editors do.

See how transcription works →

Get the next one in your inbox

Subscribe — free, twice weekly