Anthropic is rolling out an update for Claude that will allow its AI chatbot to “remember” previous conversations without asking. An update for all paid subscribers should make Claude more useful and convenient.
Anthropic says Max subscribers will be able to enable Claude’s “memory” in their settings starting today. The feature, which has been available to Team and Enterprise users since September, allows Claude to remember details from previous chats. Pro subscribers will see the storage feature “roll out in the coming days,” Anthropic said. The company hasn’t said whether it plans to make this feature available to free users in the future.
Anthropic says the goal is “full transparency.” Users will be able to clearly see what Claude remembers, rather than “vague summaries,” he said. Specific memories can also be turned on/off and edited during natural conversation. For example, you could tell Claude to focus on specific memories or to “forget about your old job completely.” Users can also create “separate memory spaces” that will separate different memories. If this works, it could be useful for stopping memories from different conversations or projects leaking into other chats, useful for separating different work projects, or using the bot for personal and professional purposes.
The highly anticipated memory feature brings Claude closer to competing chatbots such as ChatGPT and Gemini. Everyone is competing fiercely for users, and memory features are designed to encourage users to stay with the company rather than starting over with another bot. Claude lags behind its rivals in terms of memory – both OpenAI and Google implemented memory features in their chatbots last year – and only gained the ability to remember previous conversations in August. Even then, you had to specifically ask Claude to remember.
Anthropic hopes to reduce the hassle of starting over by allowing users to import memories from ChatGPT or Gemini. You will need to copy and paste them. Memories can also be exported from Claude “at any time,” it said. “No lock.”
However, the Chatbot’s memory turned out to be a source of division. Although it was hailed as a useful feature, some experts warn reminder can facilitate maintain or enhance delusional thinking and other mental health problems, commonly calledAI Psychosis”, especially considering the sycophantic tendencies of some models.