\n\n
Recent months have seen a flurry of high‑profile licensing agreements between major publishers, platforms and AI developers, signaling that the industry is moving from ad‑hoc data grabs to structured, money‑backed contracts.
Key developments
- The Associated Press struck a two‑year deal with OpenAI in July 2023 to license its archive back to 1985, trading access to OpenAI’s technology for content rights, though the financial terms remain private.
- In May 2024, News Corp announced a multi‑year deal with OpenAI covering the Wall Street Journal, New York Post, Barron's and MarketWatch. The Wall Street Journal suggested the contract could be worth over $250 million across five years, marking one of the largest disclosed sums for news‑to‑AI licensing.
- Reddit entered a partnership with Google in February 2024, licensing its user‑generated content for AI training at an estimated $60 million per year, illustrating that social platforms are also monetizing their data for AI.
- Shutterstock broadened its six‑year arrangement with OpenAI in July 2023, extending coverage to images, video, music and metadata. While the renewal’s price was not disclosed, the company’s AI licensing segment generated $104 million in 2023 and is projected to reach about $138 million in 2024 across deals with multiple AI firms.
Why it matters
These deals formalize the value of proprietary content for training and powering generative models. By attaching explicit cash and credit terms, publishers and platforms are establishing a revenue stream that could offset potential losses from AI‑driven content substitution. The disclosed figures—$250 million for news, $60 million annually for Reddit, and over $100 million in Shutterstock’s AI licensing revenue—demonstrate that AI developers are willing to pay substantial sums for high‑quality, curated data.
What to watch next
- Deal transparency: Expect more firms to disclose financial details as the market matures, allowing clearer benchmarking of data‑valuation.
- Regulatory pressure: Ongoing copyright litigation and proposals such as the DPIIT’s blanket licence model could reshape how licensing is mandated.
- Platform expansion: Google’s AI contribution pilot and similar programs may broaden to include more publishers, potentially scaling payments beyond the current handful.
- Competitive dynamics: As AI providers secure licensed data, rivals may seek alternative sources or develop in‑house corpora, influencing pricing and contract structures.
Overall, the surge in licensed agreements underscores a shift toward a regulated, revenue‑generating ecosystem for AI training data, with billions of dollars poised to flow between content owners and AI developers in the coming years. [[1]]