Anthropic to change data retention policy after customer feedback
Anthropic will abandon its controversial data retention policy following numerous responses from corporate clients and introduce a new Enterprise Frontier Safeguards solution. It is intended to give businesses more control over how data is reviewed, stored, and processed, CNBC reports.
30-day data retention
Anthropic introduced its current policy in June alongside the launch of the Claude Fable 5 and Mythos 5 models. The company required all traffic from these models to be retained for 30 days, explaining that this was necessary to combat abuse as well as sophisticated and new cyberattacks.
At the time, Anthropic said it would not use this data for purposes unrelated to security, including model training. However, some corporate users expressed concerns about this approach.
New controls for businesses
Kate Jensen, head of Anthropic’s U.S. business, said the company spent hundreds of hours working with clients on an alternative solution. According to her, Enterprise Frontier Safeguards will enable automated security monitoring without Anthropic employees reviewing the data.
More current news is available on the UA.News Telegram channel Telegram.
The company will not charge a separate fee for Enterprise Frontier Safeguards. These controls will work both for clients using Anthropic’s technology directly and for those accessing it through cloud providers.
Phased rollout
Anthropic plans to roll out Enterprise Frontier Safeguards in phases and aims to make the solution more widely available in the fall. Jensen said she expects the deployment of Anthropic’s technologies to accelerate, including for clients’ most sensitive workloads.
The company also introduced the Claude Fable 5.1 and Mythos 5.1 models, which, Anthropic says, have improved capabilities for programming, knowledge work, and research.
