Renews.
← AI

OpenAI Data Controls and Azure Retention Policies

Users can manage data sharing settings while Azure-hosted models retain flagged content without specific time limits.

AI·1 source
Photo: TheDigitalArtist / Pixabay

Users can check their data control settings at chatgpt.com/settings/data-controls. The setting labeled 'Improve the model for everyone' is located in these user data control settings. Users can turn off data sharing via a link to OpenAI's privacy policies. OpenAI models hosted on Azure have documented asynchronous filters that can flag risks cross-conversation. Microsoft contracts removed any specific retention window for flagged content. Some flagged content is retained even if filters are active.

The retention of flagged content may have contributed to the Navier-Stokes leakage incident. The Navier-Stokes leakage involved airflow design discussions. Discussions about weapons design might trigger safety flags resulting in data retention and training. Daybreak plans to start using content from personal accounts to improve models starting November 29. Anthropic settled for $1.5 billion for using pirated books in training. Major cloud providers are now described as LLM shops.

Sources

Every sentence of this story was checked against facts quoted from these articles. How we make the news

Spotted an error? Tell us

Sign in

Keep the stories you like in your favorites.

Forgot your password?

No account?