Anthropic to Adjust Data Retention Policy for Advanced Models, Enabling Enterprise Self-Hosting

According to Bloomberg, Anthropic plans to adjust its data retention policy for its most advanced AI models, giving enterprise customers greater control over their data. The company is expected to launch a new security system later this year, under which enterprise customers will still be required to retain data for 30 days, but can choose to do so on their own cloud infrastructure rather than storing it on Anthropic's servers. Sources familiar with the matter said the system has been in preparation for several months and was developed in coordination with over 100 customers in highly regulated industries. Anthropic announced in June that it would retain all customer data for 30 days for its more powerful Mythos and Fable models, as well as future frontier products, to identify and guard against new types of cyberattacks exploiting its technology, while stating it would not use the data for training. The company acknowledged in a recent report that the policy may be unpopular with customers accustomed to zero retention and could impact business. Anthropic declined to comment on the matter. Competitor OpenAI has also recently begun testing a solution called "private security processing" to avoid retaining customer data, with plans for a broader rollout in September.