Claude has quickly built up a reputation as one of the best AI chatbots on the market. However, while it shares a lot of similarities with its biggest competitor, ChatGPT, they offer very different experiences.
One of the most notable differences between the two is Anthropic’s very different approach to memory and privacy.
Ask Claude what it knows about you and you’ll be told that, unless you’ve had a long conversation already, it doesn’t retain information from previous conversations so there is no memory of you.
It will tell you that it knows your rough location (just the country) but otherwise, it is starting fresh with no knowledge.
ChatGPT, on the other hand, identifies this as one of its key selling points. Ask ChatGPT what it knows about you and it will give a rough profile of you and your needs.
So, what does this mean and is it a good or bad thing that Claude knows nothing about you?
Claude, the forgetful
This may sound like a flaw, but it is more of a design choice from the Anthropic team. Resersving memory of previous conversations has a lot of benefits, with the chatbot altering answers to better fit you, and having on-hand knowledge of key details you’ve already mentioned.
However, for the privacy-cautious, this sounds like a nightmare. Anthropic has made a point of focusing on privacy.
Claude doesn’t train on your data, and it doesn’t store any conversation history locally. This does lower the model’s ability to know you better, but the payoff is a more secure experience.
This isn’t to say there is no memory retention on Claude. Anything mentioned in a conversation can be drawn back on as long as you stay in that conversation.
You can also save and interact with Artifacts - Claude’s version of building apps with prompts. These can be saved and gone back to. In some of these cases, the apps can retain information about you if you let them.
Should your AI chatbot have memory?
Whether this is a feature that you need, or even want, will come down to how you use AI chatbots. If you like to use it as a journaling system or something that explains things in a personalized way to you, it will make a huge difference, improving the experience drastically.
However, from a privacy perspective, or even just for someone who is only using chatbots for the occasional question or query, a lack of memory can actually be a good thing, offering a fresh angle for every prompt.
The level of memory offered will depend entirely on which model you use. ChatGPT and Claude are on opposite ends of the spectrum, but plenty of others fall somewhere in the middle.
Gemini, for example, will use relevant information from previous chats to craft a response. While it doesn’t hold a distinct profile of you, it can identify key points to bring back up when necessary.
Le Chat, a less popular chatbot offers an opt-in memory scheme. This lets you get a similar experience to ChatGPT or Gemini, but only if you choose that as an option.
Memory in chatbots comes down to how much you’re willing to give away and how much you want to get back in that exchange.
More from Tom's Guide
- I’ve started using ChatGPT’s incognito mode every time — here's 4 reasons why and how to do it
- This ChatGPT prompt gives me a 10-minute productivity boost every Monday
- Still shopping blind on Amazon? This ChatGPT trick helps me avoid bad buys — and it works every time
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
