New York Times Sets Hard Criteria for AI Licensing as Content Marketplaces Lag
The Development
The New York Times VP of strategic partnerships Adam Greenberg, speaking at the Digiday Publishing Summit in Miami, outlined the publisher's framework for evaluating AI licensing partnerships and delivered a candid assessment of the current market. Greenberg, who spent 14 years at Google building ecosystem products including AMP and Web Stories, described AI content marketplaces as 'underdeveloped' and warned that real-time content valuation systems for AI training and inference are years from maturity. The Times has already pursued legal action against AI developers over unlicensed content use, making its public licensing posture a material signal — not just a policy position — for how the most litigious major publisher in the space is navigating commercial arrangements with AI companies.
Our Take
Greenberg's framing matters because the Times is not a passive commentator — it is the publisher that sued OpenAI and Microsoft and has the resources to define what acceptable AI content licensing looks like at the premium end. When its VP of strategic partnerships calls the marketplace 'underdeveloped,' that is a market-shaping statement. AI developers who want Times content are being told: build a real commercial infrastructure or expect continued legal exposure. For brands, the downstream consequence is that AI models trained predominantly on unlicensed or lower-quality content will increasingly be the alternative — and the quality gap between licensed and unlicensed training data is a risk that brand safety teams are not yet pricing in.
What Changed
Publishers with the market position of the Times can now negotiate structured licensing terms rather than accept boilerplate opt-out arrangements. This shifts content access from a crawl-and-train default to a negotiated commercial relationship — with compensation, attribution, and usage scope as standard deal terms.
Marketing Impact
Brand safety and content strategy teams face growing exposure from AI tools whose training provenance is opaque. If premium licensed content becomes concentrated inside a small number of enterprise AI agreements, marketers using general-purpose models may be operating on a lower-quality, higher-legal-risk content base without visibility into the difference.
Competitive Implication
AI developers with established publisher licensing portfolios — including those who have reached deals with News Corp, Condé Nast, and the Associated Press — gain a content quality and legal defensibility advantage. Developers without such agreements remain exposed to litigation and are building on a deteriorating unlicensed content base as major publishers harden their terms.
Strategic Outlook
Greenberg's prediction that real-time content marketplaces will take years to mature sets a credible timeline: licensing will remain bespoke and negotiated through at least 2028. Expect the Times to use that window to establish deal structures that become the template other premium publishers reference, steadily increasing the cost floor for AI content access.
The Exploit
Action Item
CMOs procuring enterprise AI platforms in Q4 2026 should require vendors to disclose their publisher licensing agreements by name — specifically whether the Times, News Corp, or AP content is included — before signing contracts that involve content generation or retrieval.