Skip to content
AI Risk ResearchIndependent monitoring and live experiments

Research content / privacy

Privacy approach

The public observatory minimizes personal data and focuses on systems, organizations and documented public-interest events.
01

Publication boundary

Private people are not named for unverified allegations. Sensitive claims require strong public-interest evidence and additional review.

02

Source discipline

Public records link to their sources but do not republish unnecessary personal details. Evidence excerpts are limited to what supports the claim.

03

Contact safety

Do not submit credentials, private keys, health information or other secrets in a correction challenge. Remove unnecessary personal identifiers before sharing evidence.

Read the full methodology