Data retention rules loosen for Anthropic's top models

Bloomberg reports Anthropic plans to let business customers keep greater control of their data when they use its most capable AI models, moving away from an earlier retention policy meant to mitigate potential cyberattacks. Anthropic expects to roll out a new safety system later this year, according to the same report. The logging requirement had put it opposite OpenAI, which is testing Private Safety Processing with early customers to pair zero data retention with a narrow safety signal.