Google is rolling out new Gemini personalisation features in the UK, including a memory setting that can use your past chats to make future answers feel more tailored. It is also adding tools that let people import memories and, over the next couple of weeks, full chat histories from other AI apps.
That sounds convenient. It also makes Gemini feel less like a blank search box and more like a long-running assistant that gradually learns what matters to you. For ordinary users, that shift is worth understanding before simply leaving every setting as it arrives.
According to Google’s UK announcement, the Memories setting is rolling out from today, with availability for all users in the coming weeks. Google says the setting is on by default, but can be turned off in Gemini app settings under Personal context, then Memory. Users can also manage and delete conversations in Gemini Apps Activity.
What Gemini memory actually means
In simple terms, memory lets an AI assistant carry useful context from one conversation to another. If you often ask about vegetarian meals, school revision, small business admin or a hobby, the assistant may use that pattern to make later answers more relevant.
That can be helpful. A chatbot that remembers your preferred tone, your regular tasks or the kind of examples you understand can save time. You may not have to repeat the same background details every time you ask for a plan, draft, summary or recommendation.
But memory also changes the relationship. Instead of each chat standing alone, your assistant may begin building a picture of you: your routines, interests, work, family context, worries, habits and preferences. Some of that may be harmless. Some may be more sensitive than people realise when they are typing quickly.
This is why the setting matters. Personalisation is not only a feature; it is a permission decision.
The import tools could make switching easier
Google is also introducing switching tools for consumer accounts. One option lets you prompt another AI app to summarise your saved preferences and personal context, then paste that summary into Gemini. Another will let you upload a ZIP file of your chat history from other AI providers, so you can search old threads and continue them inside Gemini.
For anyone who has spent months using a chatbot for recipes, work notes, travel planning, coding help, study support or family admin, that could reduce the friction of moving between AI services. It may also intensify lock-in. Once years of conversations sit inside one assistant, people may feel less free to leave, even if a rival app becomes cheaper, safer or more useful.
ManyHands has already looked at Gemini chat imports, and the practical point remains the same: before moving old conversations, look at what is actually inside them. AI chats can contain more personal detail than email inboxes because people often use them when they are thinking out loud.
What UK users should check now
If Gemini memory appears in your account, the first step is not to panic or rush to switch it off. The better habit is to make a deliberate choice.
- Open the setting yourself. In the Gemini app, check Personal context and Memory rather than assuming the default matches your preference.
- Review old chats before importing them. Look for medical questions, family disputes, workplace problems, financial worries, children’s details, passwords, addresses or anything you would not want reused later.
- Separate useful preferences from sensitive facts. “I like plain English summaries” is very different from detailed notes about a private situation.
- Delete what should not be kept. If an old conversation was only useful once, it may not need to become part of your long-term assistant context.
- Be especially careful on shared devices. A family tablet or work laptop may blur personal and professional use.
The same principle applies when chatbots help with shopping, work or admin. As we noted in our guide to AI shopping assistants, convenience is useful only if you still understand what information is being used to steer the answer.
When memory is useful
There are plenty of sensible uses for AI memory. A student might want an assistant to remember exam subjects and preferred revision style. A carer might save non-sensitive routines that make planning easier. A freelancer might want a chatbot to remember formatting preferences, brand tone or regular types of task.
For work, memory can reduce repetitive briefing, though it should be handled carefully. If your employer has rules about client data, confidential files or internal documents, do not assume a consumer AI account is the right place for them. Many organisations will prefer managed workplace tools with clearer controls.
For home use, the safest approach is selective memory. Let the assistant remember low-risk preferences that genuinely improve answers, but avoid feeding it private material just because it feels conversational. If a topic would make you uncomfortable if it resurfaced unexpectedly in a future answer, it probably does not belong in long-term memory.
The takeaway
Gemini’s UK personalisation update is part of a bigger change in consumer AI. The most popular assistants are moving from one-off answer machines towards persistent companions that can remember, anticipate and adapt.
That may make them more useful. It may also make privacy choices less obvious. A good default for users is simple: treat AI memory like a contacts book, not a bin. Keep what helps, remove what does not, and do not import years of old chats without checking what you are bringing with you.
If you use Gemini, this week is a good time to open the settings, read the memory options, and decide how personal you actually want your personal assistant to be.
