Adobe refutes allegations of AI training through customer data

Adobe refutes allegations of AI training through customer data

Adobe does not use customer projects to train generative AI models, Chief Product Officer Scott Belsky said in a statement to Bloomberg. The company denies allegations of AI training through customer data.

Generative AI models such as ChatGPT and Dall-E learn to create new materials by analyzing content from various authors. Content creators are rightfully concerned that their materials are being used for AI training without notice or compensation.

Late last year, a screenshot of Adobe’s new terms of use got leaked. The terms reportedly stipulated that Adobe would use images and videos from customers to train generative AI models. This led to an outcry.

CPO Belsky recently addressed concerns and told Bloomberg that Adobe has “never” used projects stored by customers to train AI models. According to the executive, the terms of use have clarified that Adobe analyzes data to improve functionality for decades.

If Adobe were to use customer projects for AI training, the company would carefully request consent with an opt-in setting, the CPO claimed.

Wake-up call

Adobe is busy integrating generative AI tools into its software. Various new AI tools were announced at its October 2022 user conference. Belsky said that the outcry surrounding Adobe’s terms of use has been a “wake-up call” for the company.