Google is splitting privacy controls in the Gemini app, separating chat history from AI training permissions so users can retain their history without allowing Google to train models on their data. The revised privacy settings are scheduled to roll out gradually over the coming weeks, according to coverage on November 18, 2026. The split is a response to a long-standing complaint: the old controls forced an all-or-nothing choice.
The Gemini app is separating two controls that used to be bundled. Chat history, whether the app remembers your conversations, is becoming a separate choice from AI training, whether Google can use your conversations to improve its models. Users will be able to keep their history and opt out of training, or the reverse, instead of choosing one setting that controlled both.
The change is rolling out gradually over the coming weeks, which means not every user will see it at once. Google has not published a firm completion date, and the rollout will likely reach different regions and account types at different times.
The old bundled control created a bad trade. A user who wanted the app to remember their conversations had to accept that those conversations could train models. A user who did not want their data used for training had to give up history, which made the app much less useful. The split removes that trade and lets users choose each thing on its own merits.
The change also reflects regulatory pressure. Privacy regulators have pushed for granular consent, where users choose each use of their data separately rather than accepting a bundle. A split control is easier to defend to a regulator than a bundled one, and it is easier to explain to a user.
For SEOs, the split affects how users interact with Gemini, and interaction is where citations are born. A user who keeps history is more likely to have long, multi-turn conversations with the assistant, and long conversations produce more citations. If the split encourages more users to keep history, Gemini's citation volume could grow.
For link builders, the training opt-out is the more important half. Content that users share with Gemini in conversations they have opted out of training will not feed the model, which means the model's knowledge of that content comes only from the public web. Public, citable content remains the way to influence what the model knows.
Watch for the rollout to reach your account, and check what the new settings actually say. The wording of the controls matters: a clear split is a real improvement, while a confusing one is a rebrand.
Watch also for whether Google applies the same split to other products. If history and training become separate choices across Google's AI products, the company's privacy story becomes much stronger. If the split stays Gemini-only, it looks like a one-off fix.
When the split rolls out to your account, the first thing to do is open the Gemini privacy settings and read what they actually say. The two controls should be clearly labeled, one for history and one for training, and each should explain what it does in plain language. If the wording is confusing, or if one control still affects the other, the split is not working as promised.
The second thing to do is decide what you actually want. Most users will want history on and training off, because history makes the assistant useful and training is the part that feels invasive. That combination is exactly what the split is for, and it is worth choosing deliberately rather than leaving the defaults in place. Defaults are a choice too, and they are usually the company's choice, not yours.
Sources: Gadgets 360, November 18, 2026.
Related reading on this site: AI Overviews Showed Up in Only 20% of Local Searches, and Mostly Copied the Map Pack, ChatGPT Wallet Lets AI Agents Pay Without Seeing Your Card Number, November 2026 Core Update Still Rolling Out as Volatility Peaks on November 14, Domain Authority Barely Correlates With AI Citations, Search Atlas Finds, Social Networks Are a Minor AI Search Source — Except LinkedIn and YouTube, and Semrush Rebuilds Traffic and Market Data With a New Model, and Old Numbers Will Change.
Linkbaits helps you create content that earns citations and backlinks. Start with the toolkit and build your first linkable asset this week.
Start free trial