model training data filtering
AssessTechniques
Methods to reduce personal data in training datasets.
Why it's here
Placed in Assess: 1 article(s) of evidence from 1 source(s), led by framework updates, with 0 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 5OpenAI Blog·5/6/2026framework_updateHow ChatGPT learns while protecting privacy
OpenAI explains how ChatGPT can improve from conversations while reducing the amount of personal data used in training. The company also describes controls that let users decide whether their chats help train AI models.