Urgent.News

What's breaking now, across thousands of outlets.

AI

Anthropic promises zero data retention – but customers must check it worked

More compliance-friendly stance could boost appeal of Fable

Anthropic promises zero data retention – but customers must check it worked

Anthropic has promised to provide zero data retention to enterprise customers using its Fable and Mythos models, but customers must apply for this feature when using the new versions, Fable 5.1 and Mythos 5.1. Anthropic's new service, Enterprise Frontier Safeguards (EFS), aims to ensure privacy while detecting model abuse. However, corporate customers are not automatically granted ZDR agreements; they must request them.

Thirty-day retention of inputs and outputs is standard for commercial customers using the API, unless other arrangements are made. In the past few months, Anthropic has observed substantial evidence of attempted misuse of AI models, leading to the need for short-term data retention to correlate events across accounts. Despite assuring enterprises that data retention is not for training purposes, many firms found it difficult to use models with data retention, especially in regulated industries.

To address this, Anthropic plans to release EFS this fall, which stores data in customer-controlled cloud infrastructure, shifting safety work to the enterprise. This move may disadvantage Fable in a market already hesitant to pay a premium, but it offers a compliance win for users of Fable. Mythos will continue under Anthropic's earlier policy, while EFS will be offered for various platforms, by invitation. Consumer plans remain unchanged.

Written by urgent.news from The Register's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at theregister.com →

More in AI

More from Tuesday 1 September →