AI Privacy Alert: Anthropic’s Claude And Cowork Will Remember You Unless You Say No

📊 Full opportunity report: AI Privacy Alert: Anthropic’s Claude And Cowork Will Remember You Unless You Say No on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

Anthropic has officially enabled shared memory between its two AI products, Claude and Cowork, by default, meaning that information the assistant learns about a user will now automatically carry over across both platforms unless the user actively disables this memory sharing feature. This shift in default settings significantly alters the privacy landscape for users relying on Anthropic’s tools, impacting how personal and professional data are stored and used.

According to Anthropic, the change means that memories accumulated by Claude—such as user preferences, ongoing projects, or personal details shared during conversations—are now accessible to Cowork, the company’s agentic AI designed to perform multi-step work tasks. Previously, users had to opt in to share such context; now, sharing is the default, with an opt-out option available in Claude’s settings. The company states that users can disable memory sharing if they wish to keep their data separate, but it is unclear whether this applies to specific memory categories or only as a global toggle.

The update was announced as part of Anthropic’s effort to improve user experience by reducing repetitive inputs and enabling more seamless task execution across products. However, privacy advocates express concern over the default opt-out model, which they say could lead to unintended data exposure, especially since most users tend not to change default settings. The move also coincides with broader industry trends where competitors like OpenAI and Google have expanded persistent memory features, raising questions about data retention and user control.

Implications of Default Memory Sharing for User Privacy

This change has notable implications for user privacy, as personal information shared in one product now automatically becomes accessible in another, potentially exposing sensitive data without explicit user consent. For privacy-conscious users, this default setting may lead to increased data collection and use, raising concerns about how stored memories are managed, whether they are used for training, and how easily they can be deleted. For the broader AI industry, the move underscores a shift toward more integrated, context-aware AI systems, but also highlights ongoing debates about data control and transparency in AI development.

Industry and Company Background on AI Memory Features

Anthropic’s introduction of memory capabilities for Claude aimed to enhance the user experience by allowing the assistant to retain information across conversations, such as user preferences or recurring tasks, making interactions more efficient. The company also developed Cowork as part of its expanding suite of agentic AI tools capable of executing complex workflows. Prior to this change, users had to manually enable cross-product memory sharing, which limited automatic data flow. The industry has seen a trend toward persistent memory in AI models, with competitors like OpenAI’s ChatGPT and Google’s Gemini expanding their memory features, often sparking privacy concerns. Anthropic’s recent default shift reflects a broader industry debate about balancing convenience with user control over personal data.

“Users can opt out of memory sharing between Claude and Cowork in settings.”

— Anthropic spokesperson

Unanswered Questions About Memory Sharing Details

Several key details remain unclear, including whether specific categories of memories are selectively shared, how existing memories are handled during the transition, and whether users can customize which memories are shared. It is also not confirmed if shared memories are used for training purposes or how deletions in one product affect the other. The timing of full rollout and whether administrators can control settings on managed accounts are also still uncertain, leaving some ambiguity about the scope and control of this feature.

Next Steps for Users and Industry Response

Users should review their Claude and Cowork settings this week to disable memory sharing if desired. Anthropic is expected to provide further clarification on selective sharing, retention policies, and data training practices soon. Industry observers will monitor how competitors respond to this shift, especially regarding default privacy settings and user control. Additionally, further updates from Anthropic may clarify whether shared memories can be pruned or managed more granularly, and how the company plans to address privacy concerns raised by users and privacy advocates.

Key Questions

What exactly is being shared between Claude and Cowork?

Memories such as user preferences, ongoing projects, and personal details shared during conversations will now be accessible to both products unless the user disables sharing in settings.

How do I stop Claude and Cowork from sharing memories?

Go to Claude’s settings and disable the memory-sharing option to keep your data separate across the two products.

Does this change affect data used for training?

Anthropic has not fully clarified whether shared memories will be used for model training or how deletions in one product influence the other. Further details are expected soon.

Will my existing memories transfer automatically?

It is not yet confirmed whether current stored memories will migrate automatically into Cowork or require manual management.

Are there risks associated with this default sharing?

Yes, privacy advocates warn that automatic sharing may expose personal or sensitive information without explicit user consent, especially if users do not review or change their settings.

Source: ThorstenMeyerAI.com