Adobe denies rumours that it trains AI generators on user images.

Adobe Chief Product Officer Scott Belsky said the company has never used user content to train a generative AI. Bloomberg reports.

“We have never used anything from our storage to train a generative AI model. Not once,” said Belsky.

Earlier, developers of the independent Krita Foundation discovered in Adobe’s terms of service a clause stating the use of AI to analyse content. Some artists interpreted this as the company using images to train generators such as DALL-E 2 and Midjourney.

Belsky described the criticism as a troubling wake-up call and said the policy is not intended to generate content. According to him, the company uses AI to analyse products and improve their features.

“We are rolling out a new evolution of this policy, which is more specific,” said Belsky.

He added that if the company ever decides to develop generative AI, Adobe will announce it publicly.

“We must be very clear about these things,” concluded Belsky.

Adobe has previously developed several products and tools that use AI algorithms. Among them is Adobe Sensei, which integrates with Photoshop, Lightroom and Premiere Pro. The tool allows creators to automate routine or repetitive tasks such as identifying faces and buildings in photos and videos.

At Adobe Max 2022, the company also showcased several Sneaks projects with artificial intelligence.

In January, artists filed suit against Stability AI and Midjourney over training algorithms on their images.

Subscribe to ForkLog’s news on Telegram: ForkLog AI — all the news from the AI world!