Adobe tells Congress to give artists the right to stop AI training on their creative work

A senior Adobe representative told Congress on Wednesday that the company supports giving artists the right to prevent AI systems from being trained on their creative work and indicated that it would support a new Congressional bill on that issue.

In a Senate Judiciary Committee subcommittee hearing on the link between artificial intelligence and intellectual property rights, Adobe general counsel and chief trust officer Dana Rao told senators that allowing AI systems to train on large sets of data is critical to ensure accurate AI results.

But Rao said Adobe, the imaging and design software giant, recognizes that boundaries need to be placed to protect creators. For this reason, he said Adobe trained its generative AI system, Firefly, on licensed images from the company’s stock photo library and other openly licensed content.

SENATORS LEAVE BRIEFING TO CONFIDENT BUT DEFECTIVE ABOUT CHINA’S “EXISTENTIAL” THREAT

Adobe told Sens. Chris Coons, D-Del., left, and Thom Tillis, RN.C., this week that artists need new protections from AI-generated content. (Getty Images)

“This approach supports creators and customers by training on the dataset that is designed to be commercially safe,” he said, adding that government has a role to play in ensuring other companies respect creative works in the same way.

“We believe creators should be able to attach a ‘do not train’ tag to their work,” Rao said. “With support from industry and government, we can ensure callers of AI data read and respect this tag, giving creators the ability to keep their data out of AI training datasets.”

Rao was asked by subcommittee chair Chris Coons, D-Del., if Congress should step in and impose an option so artists can ensure their work cannot be used by AI systems for education some data.

“I think there’s an opportunity for Congress to mandate that a tag like that, a credential like that, be carried wherever the content goes,” he said.

Rao went further by saying that Congress should create a federal law that prevents AI systems from unfairly impersonating artists’ styles.

HOW AI SHAPED THE VITAL PRESIDENTIAL ELECTION OF NATO ALLIES

An Adobe representative told Congress that the company supports giving artists the right to prevent AI systems from being trained on their creative work. (Rafael Henrique / SOPA Images / LightRocket via Getty Images / File / Getty Images)

“We believe artists should be protected from this type of economic harm, and we propose that Congress institute a new federal anti-impersonation law that gives artists the right to sue against someone who intentionally attempts to impersonate their style or their likeness,” he said.

“Holding people accountable for abusing AI tools is a solution we believe goes to the heart of some of our customers’ problems, and this new right would help address that concern,” Rao added.

Lawmakers also heard on Wednesday from an artist who has worked on feature films, who said artists could use new rules related to copyright protection from AI systems.

“I’ve never worried about my future as an artist until now,” Karla Ortiz told the subcommittee. “Generative AI is unlike any previous technology. It is a technology that uniquely consumes and exploits the hard work, creativity and innovation of others.”

GENERATIVE AI TOOLS LEAD TO GROWING DEEPFAKE FRAUD

Congress has been scrambling for major new legislation to regulate AI, but it hasn’t gotten very far yet. (Reuters / Dado Ruvic / Files / Reuters Photos)

“I found that nearly all of my work, the work of nearly every artist I know, and the work of hundreds of thousands of artists had been taken without our consent, credit or compensation,” he said. “These jobs have been stolen and used to train for-profit technologies with datasets that contain billions of data pairs of images and text.”

GET FOX BUSINESS ON THE GO BY CLICKING HERE

Members of Congress are talking about a variety of ideas for how to regulate AI, but no major legislation has gotten off the ground yet. Earlier this year there was talk of establishing a new federal agency to manage AI, or at least a commission, but none of these ideas have gone anywhere.