A display deal that tells us where AI search is headed
The Getty Images OpenAI deal is a multi-year AI image licensing agreement that lets ChatGPT Search images draw from Getty’s library for on-screen display while keeping model-training rights and financial terms undisclosed, marking a test case for how licensed photos and AI search can coexist in the same interface.
This is not a minor integration; it is a statement about where AI search is going. Getty Images and OpenAI have signed a multi-year display partnership to bring Getty’s licensed content into search and discovery experiences inside ChatGPT. In practice, that means professional photos will appear directly in AI answers instead of sending users off to separate stock sites. The headline: AI platforms are finally paying—at least someone—for imagery, and markets are rewarding that move far faster than courts ever did. But the deal is also a warning: if creators are not named in these contracts, they risk being treated as line items rather than partners in the new AI image economy.
Markets cheer the AI image licensing agreement—maybe too loudly
Investors did not wait for the fine print. On news of the Getty Images OpenAI deal, Getty’s stock exploded in premarket trading, at one point more than tripling from its prior close, before settling to a gain of about 123% at USD 1.35 (approx. RM6.20) per share by Monday’s open. Another report clocked the intraday jump at roughly 108%, as shares climbed by 65 cents to USD 1.26 (approx. RM5.80). For a stock that had lost more than half its value this year, that is not a bounce; it is a vote for a new business model.
The market message is blunt: if AI pays for content, investors believe there is life left in old media catalogs. Licensing, not litigation, is what traders want to see. Having tested the courtroom route against another AI developer and largely lost, Getty appears to have concluded that licensing is the surer path to getting paid. The risk is that this optimism is priced on opacity. No one outside the two companies knows the size of the check, the revenue split, or how long the money lasts—and yet billions in market value are suddenly being repriced on faith.
What changes for users: richer answers, more labels, new biases
For ordinary users, the upside is clear: more useful visuals, fewer dead-end links. Getty Images says its licensed content will appear in ChatGPT, improving the richness of visual responses on the platform. The deal will see images from Getty’s library surface within ChatGPT’s search display, adding richer visual features to the chatbot. OpenAI can now show professional photos and editorial images directly inside ChatGPT search answers.
But the quality upgrade comes with subtle shifts in power. When ChatGPT Search images are drawn from a single licensed channel, that channel gains visibility and brand weight. Display rights determine what users see; training rights determine what the AI silently learns from. If answers prioritize paid libraries, free or open alternatives risk being buried. Users will rely on visible credits and source links to understand which images are licensed and from whom. This is good for transparency, but it also normalizes AI as a commerce layer: every image is both an answer and an ad for the catalog behind it.
Creator rights in an AI partnership built on secrecy
The most troubling part of this creator rights AI partnership is what we still do not know. Neither Getty nor OpenAI has disclosed the financial terms. Getty has also not said whether the arrangement includes using images to teach or tune AI systems. Commercial terms remain private, including pricing, contributor economics, or any separate training-rights payment. That is a gaping hole, given that Getty works with nearly 600,000 content creators and hundreds of content partners, covering more than 160,000 events each year.
Creators are being told, in effect: trust the agency that already lost one AI lawsuit to negotiate your future. Getty’s legal history, including its 2023 suit against another image-generation company over alleged scraping of 12 million photos, keeps the training question prominent. The current deal explicitly stops short of letting OpenAI train its DALL·E generator on Getty’s archive. That sounds protective, but it also locks creators into a system where display rights are monetized while the most valuable long-term asset—training access—sits on a separate, undisclosed track that they cannot see.
A template for media–AI deals that still separates display from training
The real precedent here is architectural: this agreement draws a bright line between what AI shows and what AI learns. Agreement terms let ChatGPT show Getty imagery while keeping the public scope focused on display rather than model training. The contract does not allow Getty images to be used to train OpenAI’s DALL·E image generator. Display rights govern what a user sees; training rights would govern whether image libraries help build or tune the underlying AI system.
Across publishing, music, and now stock photography, licensing deals have become the media industry’s preferred answer to the AI question. OpenAI’s deal extends Getty’s controlled-licensing posture into consumer AI search rather than marking a broad surrender of rights. That is progress: it shows traditional media companies can partner with AI platforms without giving up everything. But it is also incomplete. If this structure spreads, we could end up with AI models that display licensed images while still being trained on unlicensed data elsewhere. The conclusion is uncomfortable but clear: until training access is negotiated and paid for in the open, AI’s relationship with creative work will remain half-legal, half-shadow economy—and this deal, for all its promise, does not change that yet.
