Adobe says it won’t train AI using artists’ work. Creatives don’t agree


When users first learned about Adobe’s new terms of service (which were quietly updated in February), there was an uproar. Adobe told users that it could access their content “in both automated and manual ways” and “use technologies such as machine learning to improve (Adobe’s) services and software.” Many interpreted this update as the company forcing users to give it unlimited access to their work for the purpose of training Adobe’s generative AI: Firefly.

Late Tuesday, Adobe issued a clarification: In an updated version of its terms of service agreement, it pledged not to train AI on its user content stored locally or in the cloud and gave users the option to opt out of content analysis.

Caught in a tangle of intellectual property lawsuits, the vague language used to first update the terms highlighted a climate of intense suspicion among artists, many of whom depend on Adobe for their work. “They’ve already broken our trust,” says John Lam, a senior storyboard artist at Riot Games, describing how award-winning artist Brian Kessinger discovered that images generated in the style of his art were being sold under his name on his stock image site without his consent. Earlier this month, the estate of the late photographer Ansel Adams publicly slammed Adobe for selling generative AI imitations of his work.

When artists began protesting, Adobe’s chief strategy officer Scott Belsky had tried to ease concerns, clarifying that machine learning referred to the company’s non-generative AI tools—Photoshop’s “Content Aware Fill” tool, which allows users to seamlessly remove objects in an image, is just one of many tools made possible through machine learning. But while Adobe insists that the updated terms do not give the company ownership of content and that they would never use user content to train Firefly, the misunderstanding has sparked a larger discussion about the company’s monopoly of the market and how such a change could threaten artists’ livelihoods at any time. Lam is among the artists who still believe that despite Adobe’s clarification, the company will use work created on its platform to train Firefly without the creator’s consent

Concerns about the non-consensual use and monetisation of copyrighted work by generative AI models are nothing new. Early last year, artist Carla Ortiz was able to prompt images of her work using her name on various generative AI models; an offence that sparked a class action lawsuit against MidJourney, DeviantArt and Stability AI. Ortiz wasn’t alone – Polish fantasy artist Greg Rutkowski found that his name was one of the most used prompts in Stable Diffusion when the tool first launched in 2022.

As the owner of Photoshop and the creator of PDF, Adobe has reigned as the industry standard for over 30 years, empowering much of the creative class. An attempt to acquire product design company Figma was blocked and abandoned in 2023 due to antitrust concerns that would prove its size.

Adobe has clarified that Firefly is “ethically trained” on Adobe Stock, but longtime stock image contributor Eric Urquhart insists that “there was nothing ethical about the way Adobe trained the AI ​​for Firefly”, pointing out that Adobe does not own the rights to any of the individual contributors’ images. Urquhart had originally placed his images on stock image site Fotolia, where he had agreed to licensing terms that specified no use for the generative AI. Fotolia was acquired by Adobe in 2015, which released some silent updates to the terms of service that later allowed the company to train Firefly using Eric’s photos without his explicit consent: “The language in the current change to the TOS is very similar to what I saw in the Adobe Stock TOS.”


Please enter your comment!
Please enter your name here

Share post:




More like this

Today’s NYT Connection Hints, Answers & Help June 21, #376

Need answers to the June 21 New York Times...

We’re still waiting for the next big leap in AI

When OpenAI announced its latest big language model, GPT-4,...

US bans sale of Kaspersky software from Russia citing security risk

The US government announced on Thursday that it is...