Google Gemini Adds Memory and Privacy Tools That Could Rival ChatGPT

Google has just rolled out two major upgrades to its Gemini AI chatbot — and they’re free for everyone, including non-paying users. The updates include a memory system that remembers past conversations for more personalized answers and a Temporary Chat mode for private, non-tracked discussions. Alongside these, Google has also introduced fresh data control settings that give users more say over what the AI can store and use.

While AI companies often chase cutting-edge features in new model releases, these updates focus on making Gemini more practical and human-friendly. The goal is to make interactions smoother, reduce the repetitive effort of explaining yourself every time, and give people better control over their privacy — an approach that mirrors similar moves by ChatGPT earlier this year.

the Original

Google announced that Gemini can now reference past chats to deliver more tailored responses. For instance, if you’ve previously told it you own a small dog, Gemini can use that detail in later advice without you re-explaining. The feature, similar to ChatGPT’s Memory, will roll out in phases, starting with the 2.5 Pro model in certain regions before expanding to 2.5 Flash.

Importantly, the memory function is on by default, but users can disable it via the Gemini app settings. You can also delete specific interactions from your activity history, ensuring unwanted data is erased.

Google is also launching Temporary Chat, which works like Chrome’s Incognito mode — conversations won’t be saved, won’t appear in chat history, and won’t be used to train AI models. However, this also means they won’t contribute to personalized responses in future conversations.

In addition, the company is rebranding its “Gemini Apps Activity” setting to “Keep Activity”. Starting September 2, if this setting is enabled, a sample of your uploads may be used to improve Google services. Privacy-conscious users can toggle it off or use Temporary Chat instead. Notably, if you already have activity tracking disabled, it will remain off after the rebrand.

Overall, the update blends personalization with stronger privacy controls, giving users more flexibility in how they interact with Gemini.

What Undercode Say:

Google’s move here is a strategic response to two competing forces in AI: the demand for deeply personalized experiences and the growing mistrust over how AI companies handle user data. By offering both — a memory that recalls details for richer interactions, and a privacy-first mode that forgets everything — Google is trying to please both sides of the debate.

From a usability standpoint, memory is a game-changer for anyone using AI for long-term projects, customer interactions, or complex problem-solving. Repeatedly feeding the same context is exhausting and breaks conversational flow. A persistent memory fixes that, making Gemini feel more like a personal assistant than a generic chatbot.

But there’s a trade-off: every personalization feature is a double-edged sword. To remember details, Gemini must store them — and stored data is always a potential privacy risk. Google is clearly aware of this, which is why Temporary Chat exists. It’s not just about user trust; it’s also a legal shield in an era of tightening AI data regulations.

This update also hints at a subtle competitive shift. OpenAI’s ChatGPT memory feature rolled out earlier this year, and Anthropic’s Claude followed suit. Google joining the race signals that persistent AI memory is quickly becoming a standard, not a premium extra. The fact that these tools are free shows Google’s urgency to keep Gemini relevant against ChatGPT, Claude, and the fast-growing Perplexity AI.

The new “Keep Activity” branding feels like more than a cosmetic change — it reframes how users perceive data retention. “Apps Activity” sounded technical and opaque; “Keep Activity” makes it explicit that you’re keeping data, and by implication, so is Google. The September 2 date is a gentle push for users to review their privacy settings now.

Interestingly, Google is rolling these updates out in staggered phases, starting with Pro users. This could be a way to test how memory interacts with different workloads before unleashing it on the broader base. It’s also likely that Gemini’s memory will eventually sync across devices and apps, making it feel like a persistent Google brain that follows you wherever you sign in.

If done right, this could transform how people perceive AI assistants — not as disposable Q\&A tools, but as evolving companions that adapt to you over months or even years. However, Google will have to tread carefully. If even a small privacy mishap happens, users will remember that more vividly than any AI memory feature.

🔍 Fact Checker Results

✅ Google has officially confirmed memory and Temporary Chat are rolling out to all users.
✅ The “Keep Activity” rebrand is scheduled for September 2, with data use opt-out options.
❌ There is no evidence yet that memory syncs across all Google apps — rollout details remain partial.

📊 Prediction

Within the next 12 months, AI assistants with persistent, customizable memory will become the default expectation for users, not a novelty. Google will likely integrate Gemini’s memory deeper into its ecosystem — including Gmail, Docs, and Calendar — giving it a massive advantage in context-aware productivity. However, the true battleground will be trust: whichever company convinces users that their personal data is safe will dominate the AI assistant race.

Do you want me to also add a comparison table showing how Gemini’s new features stack against ChatGPT and Claude? That would make the analysis even sharper.