Adobe’s terms of service update sparks fears over AI training

Adobe’s recent notification about terms of service update stirred online outrage among users, particularly regarding the company’s plans to access users’ content “through both automated and manual methods.”

Many Adobe customers were worried that the Photoshop maker would use their content to train artificial intelligence.

Several users expressed concerns about privacy infringement and exploitation of their creative works without consent, fearing that their unique styles and techniques could be replicated by AI without permission or compensation.

At the core of the controversy are the changes that were slipped into the ToS earlier this year, which would allow Adobe to employ users’ content for various purposes, including training its AI models, similar to the Windows 11 Recall feature.

Many grew concerned that Adobe would train Firefly, its generative AI model, on user content.

Adobe's clarification
In response to the uproar, Adobe clarified that it doesn't train Firefly on customer content and respects user ownership rights.

Adobe emphasized that Firefly’s AI models train on licensed content, such as Adobe Stock and public domain material. However, concerns persisted due to the perceived ambiguity in the language of the notification.

Scott Belsky's Statement
Adobe’s Chief Product Officer, Scott Belsky, acknowledged that the wording in the notification is unclear, emphasizing that “trust and transparency couldn’t be more crucial these days.”

The incident echoes broader anxieties surrounding tech firms’ use of user-generated content without explicit consent. 

For instance, Google faced backlash for using website data to train its AI without proper permissions and Facebook was criticized for utilizing user photos to improve its facial recognition technology without clear consent. And who can forget the most recent Scarlett Johansson-OpenAI tiff over the tech company’s voice assistant that sounded eerily like her in the movie Her?