Two events on the same December day captured the two directions the AI-and-publishing standoff can take: as one major tech vendor struck licensing deals with publishers, another was sued by the most influential publisher in the United States. On December 5, 2025, The New York Times filed a copyright lawsuit against AI search startup Perplexity — the same day Meta announced content deals with a roster of major news organisations.
The Times’ case against Perplexity
Filed in the Southern District of New York, the suit accuses Perplexity of unlawfully scraping the Times’ articles, videos and podcasts and repackaging near-verbatim copies as its own output. The complaint also raises a trademark angle: the Times argues Perplexity harms its brand by attributing AI “hallucinations” — fabricated answers — to Times journalism.
Times spokesman Graham James said the publisher believes in responsible AI development but strongly objects to “Perplexity’s unlicensed use of our content to develop and promote its products.” Perplexity’s head of communications, Jesse Dwyer, countered that publishers have been suing new technology companies for a century — radio, TV, the internet, social media and now AI — and argued those efforts have never succeeded.
The dispute is not new. The Times sent Perplexity a cease-and-desist notice back in 2024 demanding it stop using Times content; the lawsuit escalates that fight. A day earlier, the Chicago Tribune filed its own copyright suit against Perplexity on similar grounds.
A growing pile of AI copyright battles
The case joins a crowded docket. The Times’ earlier lawsuit against OpenAI and Microsoft, filed over the use of its archive for model training, is still ongoing. Getty Images sued Stability AI; a group of authors sued Anthropic, which settled in 2025 and agreed to pay $1.5 billion. Perplexity itself faces additional pressure: Reddit sued the company in October 2025 over data scraping, and in November Amazon moved against Perplexity’s shopping agent, alleging it made purchases on Amazon’s site while posing as a human user.
Legal observers note the central questions are factual ones: how the models were trained, what the products actually reproduce, and whether AI answers displace the market for original journalism. Fair use — the legal doctrine permitting some use of copyrighted material — will be the key battleground, and outcomes may differ between training a model and republishing retrieved content.
Meta takes the licensing route
Meta’s same-day announcement showed the alternative path. The company signed multi-year licensing agreements with publishers including CNN, Fox News, Fox Sports, USA Today, France’s Le Monde, People Inc., The Daily Caller and The Washington Examiner. Content from these partners feeds real-time news answers in the Meta AI assistant, with links back to the publishers’ sites. Financial terms were not disclosed.
The move is a notable reversal: Meta spent recent years pulling away from paying for news on its social platforms, and is now compensating publishers again because AI assistants need verified, current information to be useful.
What it means
Litigation and licensing are two answers to the same unresolved question: what is AI access to journalism worth, and who pays for it? Several observers argue case-by-case lawsuits are impractical and that the industry will eventually need standardised licensing structures, similar to those developed in the music business. Until courts rule or such structures emerge, every AI company that touches news content is choosing between Meta’s route and Perplexity’s risk.
What to watch
Key milestones ahead: Perplexity’s formal response and any motion to dismiss, progress in the Times–OpenAI case (whose outcome will shape every suit that followed), whether more regional publishers follow the Chicago Tribune, and whether Meta’s deals produce measurable traffic for its partners — the test of whether licensing actually pays. For how these dynamics affect smaller firms building on AI, see why AI agent projects stall without governance.
