Adobe Creative Cloud uses customer content to train AI systems

Occurred: January 2023

Can you improve this page?
Share your insights with us

Adobe has been automatically analysing customer content stored on Creative Cloud to train its AI algorithms, according to media reports.

The discovery sparked graphic designers, artists and other customers to share their concerns that Adobe is abusing their privacy and stealing their work to improve its own automated systems. Some saw it as another indication that their jobs are at risk of being replaced by robots.

Adobe's content analysis FAQ states it may use machine learning 'to develop and improve our products and services' and 'provide product features and customize our products and services ', providing examples such as the correction of perspective in images and automatically enhancing a document's headings and tables.

The row reflects broader concerns amongst artists, illustrators and others that their work is being scraped and used to train generative AI models such as DALL-E and Midjourney without their consent, thereby abusing their IP, commoditising their output, and potentially putting them out of work.

Adobe responded by saying it 'does not use data stored on customers’ Creative Cloud accounts to train its experimental Generative AI features.' In a Bloomberg interview, Adobe Scott Belsky later claimed the company never trained its generative AI services on customer projects.

Launched in 2011, Adobe Creative Cloud is a set of 20+ Adobe graphic design, video editing, web development, and photography applications and services delivered over the Internet.

Operator: Adobe users
Developer: Adobe
Country: USA
Sector: Business/professional services
Purpose: Improve products, services
Technology: Machine learning; Pattern recognition; Object recognition
Issue: Privacy; Confidentiality; Employment
Transparency: Governance; Marketing; Privacy

Page info
Type: Incident
Published: January 2023
Last updated: February 2024