Claude Gains Memory to Rival ChatGPT and Gemini
▼ Summary
– Anthropic is rolling out a memory feature for Claude that allows the AI chatbot to remember details from past conversations for all paid subscribers.
– The memory feature is now available to Max subscribers and will roll out to Pro users soon, having previously been accessible to Team and Enterprise users since September.
– Users can view, edit, toggle, or delete specific memories and create distinct memory spaces to separate different conversations or projects.
– This update brings Claude’s capabilities closer to rivals like ChatGPT and Gemini, with the ability to import and export memories to avoid user lock-in.
– While memory is seen as useful, some experts warn it could sustain delusional thinking or “AI psychosis” due to models’ sycophantic tendencies.
Anthropic has launched a significant update for its Claude AI chatbot, introducing a memory feature that allows the system to retain details from previous conversations automatically. This enhancement, now available to all paid subscribers, is designed to make interactions with Claude more seamless and personalized, eliminating the need for users to repeatedly provide the same context.
Starting immediately, Claude Max subscribers can activate the memory function through their account settings. Team and Enterprise users have had access since September, while Pro subscribers will see the feature become available in the coming days. Anthropic has not indicated whether free users will eventually gain access to this capability.
The company emphasizes that the memory system operates with complete transparency, enabling users to view precisely what Claude recalls rather than encountering ambiguous summaries. Individuals can manage these memories through natural dialogue, activating, deactivating, or editing them as needed. For instance, you might instruct Claude to prioritize certain memories or completely disregard previous employment details. Additionally, users can establish separate memory spaces to prevent information from one project or conversation from overlapping into another. This proves especially valuable for those using Claude across both personal and professional contexts.
Claude will proactively inform users about the specific details it has stored, ensuring clarity and control. This long-awaited memory capability narrows the gap between Claude and competing chatbots like ChatGPT and Gemini, as all major platforms vie for user loyalty through increasingly sophisticated features. While OpenAI and Google implemented memory functions last year, Claude only began retaining conversation history this past August, and initially required explicit user commands to do so.
To ease the transition for new users, Anthropic enables the importation of memories from rival platforms such as ChatGPT or Gemini, though this currently involves manual copy-pasting. The company also highlights that memories can be exported from Claude at any time, underscoring a no lock-in policy.
Despite its advantages, chatbot memory remains a contentious topic. Although many praise its utility, certain experts caution that persistent recall could reinforce or intensify delusional thought patterns and other psychological issues, sometimes referred to informally as “AI psychosis.” This risk is compounded by the observed tendency of some AI models to exhibit overly agreeable or sycophantic behavior.
(Source: The Verge)
