The company announced the change to users by email on October 19. Matching revisions have been published for the UK, Australia, and Canada; the Australian text contains additional detail on automated account restrictions and financial controls.
The new policy explicitly says automated systems, with human review where appropriate, may assess messages, listings, images, videos, livestreams, comments, transcripts, captions, reports, and other user-generated content. eBay says enforcement can include removing or restricting content, limiting buying and selling functions, or blocking an account. The policy also permits eBay to limit detail about its detection, moderation, and enforcement systems.
On AI, eBay says it primarily uses publicly available information to train generative models, but may use personal data from marketplace activity when it considers that necessary and proportionate. Its help material lists listings, product descriptions, images, item attributes, browsing and purchase history, support interactions, and feedback among relevant data categories. The source does not establish that the November wording introduces new monitoring tools or new data uses, rather than disclosing existing practices more specifically. It also does not establish that AI-training opt-out is available to US users.