Anthropic’s AI safety report exposed that the company can access full user logs, including prompts, questions, and uploaded files, raising serious confidentiality concerns for enterprise customers. As a result, organizations such as Nvidia, Palantir, Booz Allen Hamilton, and an unnamed utility company have restricted or rejected Claude use due to Anthropic’s data retention policies. #Anthropic #Claude #Nvidia #Palantir #BoozAllenHamilton
Keypoints
- Anthropic’s safety report revealed it can read complete user logs.
- The report highlighted risks involving hacking, surveillance, fraud, and sensitive data circulation.
- Nvidia limits Claude to low-sensitivity tasks and uses Nemotron for core operations.
- Palantir requires an irrevocable zero-data-retention commitment before using Claude.
- Some enterprises and utilities have canceled or restricted Claude use over data retention concerns.
Read More: https://securityonline.info/anthropic-data-privacy-concerns/