AI Did Not Eliminate Copyright. It Started Repricing It.

Getty Images looked like a company trapped on the wrong side of generative AI.

Its traditional business was easy to understand. Photographers and agencies supplied images, while media companies, publishers, brands, and advertisers paid for licenses. Getty helped customers solve a difficult problem: whether an image could be used commercially and who owned the rights.

Then image generation became cheap.

A user who once searched a stock library for a futuristic city or a business meeting could ask a model to produce something similar in seconds. The market began to assume that generic stock imagery would lose value, and Getty's share price reflected that fear.

By June 19, 2026, Getty shares had closed at USD 0.61. The company had also received a New York Stock Exchange compliance warning after trading below one dollar for an extended period.

Then the story changed.

The OpenAI Agreement Changed Getty's Role

Getty announced a partnership that would make its licensed image library available in ChatGPT search and discovery experiences, with source and rights information attached.

The public description pointed to display and discovery, not authorization to train DALL-E on Getty's archive. That distinction matters.

A conversational search product cannot treat every visual as decoration. When a user asks about a real news event, a public figure, a sporting moment, or a historical incident, a synthetic image may look convincing while showing something that never happened.

Images carry an evidentiary quality. People often assume that a photograph proves a scene existed.

Getty's archive therefore offers something a model cannot simply generate on demand: authentic images connected to a time, place, creator, agency, and licensing record.

From Stock Library to Trust Infrastructure

The market reaction was dramatic. Bloomberg reported that Getty shares rose as much as roughly 200% in pre-market trading after the partnership became public.

The percentage move was amplified by the very low starting price, and short covering or speculative trading may have contributed. The more durable question was whether Getty's role had changed.

If Getty is only a website selling generic image downloads, generative AI is a direct substitute. Users can create many ordinary illustrations cheaply, and a large part of the low-end stock market may continue to face price pressure.

But Getty also owns and represents a different type of asset: real-world visual records with provenance and legal permissions.

That asset can become more important as AI search expands. A model can create an attractive picture, but it cannot create an authentic photograph from an event that already happened. It also cannot invent a clean chain of rights after the fact.

The gap is trust.

Abundant Content Makes Verification More Valuable

The internet has faced this problem before. Search engines and social platforms made distribution cheap, but the value of professional reporting, reliable archives, and licensed media did not disappear. It changed shape.

Generative AI pushes that process further because it does not only copy or distribute content. It can create endless new content at almost zero marginal cost.

When expression becomes abundant, generic expression is likely to become cheaper. At the same time, material that is authentic, traceable, licensed, and usable inside commercial or legal workflows may become more valuable.

AI does not reduce every piece of content to zero. It separates content into different economic layers.

One layer is cheap, synthetic, and useful for low-risk illustration.

Another layer carries provenance, rights, accountability, and evidence.

The first layer is increasingly commoditized. The second may gain bargaining power.

Why Search Is Different From Image Generation

The partnership also signals that AI products are becoming search and discovery interfaces.

A pure image generator can create an imagined scene. A search product has to connect users with reality. It needs sources, citations, permissions, and a way to show where information came from.

That means the commercial value around AI may not stay concentrated inside the model. It may spread to the less glamorous infrastructure surrounding it: licensed archives, identity, provenance, rights management, verification, and liability.

These functions are expensive precisely because someone has to take responsibility for them.

Getty Is Not Automatically Rescued

One partnership does not resolve Getty's balance-sheet, competitive, or listing risks. The economics of the agreement were not fully disclosed, and public information did not establish a broad model-training license.

Getty still has to prove that it can turn archive access, provenance, and licensing into recurring value inside AI products. It also has to defend the parts of its traditional business that are most exposed to synthetic alternatives.

The agreement is therefore better understood as a route back into the value chain, not a complete turnaround.

Practical Takeaway

The deeper question for every content company is simple: what does it own that cannot be reproduced by typing another prompt?

For Getty, the answer may be less about visual beauty and more about reality, provenance, and permission.

AI did not make copyright irrelevant. It made generic content cheaper and forced trusted content to justify a new price.

Sources include Bloomberg, The Wall Street Journal, The Times, and public Getty/OpenAI cooperation details cited in the original Chinese analysis. This article is for general information only and is not investment advice.