Anthropic Merges Claude, Cowork Memory Across 3 Plans as Shared Context Raises Privacy Tradeoffs
Updated
Updated · ZDNet · Aug 25
Anthropic Merges Claude, Cowork Memory Across 3 Plans as Shared Context Raises Privacy Tradeoffs
3 articles · Updated · ZDNet · Aug 25
Summary
Anthropic has made Claude chat and Claude Cowork use one two-way memory system, so context built in either product now carries into the other unless users opt out.
Memory now updates during conversations rather than after they end, and Anthropic says existing chat history can inform Cowork tasks immediately, reducing the need to rebrief the AI.
Free, Pro and Max users get the feature on by default across web, desktop and mobile, while Teams and Enterprise accounts have memory and sensitive-topic storage off by default.
Sensitive topics such as health, politics, religion and gender identity are excluded from memory by default, though users can allow some of them; Anthropic says items like ID numbers, criminal history and immigration status are never stored.
The change fits a broader push to consolidate AI features into cross-product "super apps," while sharpening concerns over how much personal context AI assistants should retain.
Anthropic's Claude now shares memories across apps to boost productivity, but could this seamless integration accidentally leak personal secrets into work tasks?
Claude's continuous memory tracks your habits with citations, but does hitting delete truly purge your data from Anthropic's backend servers?
If an AI constantly remembers your past preferences, could its persistent memory eventually trap your workflows in a rigid creative echo chamber?