Incident #A30E9F · 09 JUL 2026minor
Facebook's AI page now names external developers' products using Meta's free models as a training source
The AI at Meta text on training data was reorganised into a list of sources. The one new source is information from products and services built by external developers using Meta's free models through APIs. Posts, photos, captions, AI interactions and public internet content were already named as training data, and the line that personal information is not specifically linked to a Meta account was kept. The change is not in a litigation section; the '[For litigation]' label is only a diff location marker.
Taken from you
- Information from products built by external developers using Meta's free models via APIs is now named as a training source
Won back
- The list of AI training sources now includes information from products and services built by external developers using Meta's free models through APIs.
Evidence · 4 of 4 quotes matched in source
Information shared on Meta Products, including things such as posts or photos and their captions, and your interactions with features that are part of AI at Meta+ inserted · ✓ matched in source
Information from third parties that provide content, including information from products and services built by external developers using our free models provided through APIs+ inserted · ✓ matched in source
When we collect training data to train our models, it may include personal information.+ inserted · ✓ matched in source
Archived version ↗ · 213 words changed