Pulse News

**Google's Gemini AI can now see your search history**

RYAN WHITWAM - MAR 13, 2025 4:28 PM

Google is continuing its quest to get more people to use Gemini, and it's doing that by giving away even more AI computing. Today, Google is releasing a raft of improvements for the Gemini 2.0 models, and as part of that upgrade, some of the AI's most advanced features are now available to free users. You'll be able to use the improved Deep Research to get in-depth information on a topic, and Google's newest reasoning model can peruse your search history to improve its understanding of you as a person. What could go wrong?

Like most big AI players, Google has a number of different models available. Gemini 2.0 Flash Thinking Experimental is the company's most capable multistep reasoning model, which can consider complex topics and gives you a window into its "thought" process. Google is adding a lot to this model in its latest round of updates, enabling a much larger 1-million-token context window, file uploads, and faster output. It also supports more Google apps with connections to Calendar, Notes, Tasks, and Photos.

With the aim of making Gemini more personal to you, Google is also plugging Flash Thinking Experimental into a new source of data: your search history. Google stresses that you have to opt in to this feature, and it can be disabled at any time. Gemini will even display a banner to remind you it's connected to your search history so you don't forget. If you grant access, the AI can allegedly understand you better and offer more relevant recommendations. It feels a bit strange to turn Gemini loose on such personal data, but Google already knows what you look up on the Internet. You're not giving up much more if you let the robot have a peek. This is apparently just the start of Google's efforts to personalize the AI.

Gemini 2.0 is also coming to Deep Research, Google's AI tool that creates detailed reports on a topic or question. This tool browses the web on your behalf, taking its time to assemble its responses. The new Gemini 2.0-based version will show more of its work as it gathers data, and Google claims the final product will be of higher quality.

You don't have to take Google's word on this—you can try it for yourself, even if you don't pay for advanced AI features. Google is making Deep Research free, but it's not unlimited. The company says everyone will be able to try Deep Research "a few times a month" at no cost. That's all the detail we're getting, so don't go crazy with Deep Research right away.

Lastly, Google is also rolling out Gems to free accounts. Gems are like custom chatbots you can set up with a specific task in mind. Google has some defaults like Learning Coach and Brainstormer, but you can get creative and make just about anything (within the limits prescribed by Google LLC and applicable laws).

Some of the newly free features require a lot of inference processing, which is not cheap. Making its most expensive models free, even on a limited basis, will undoubtedly increase Google's AI losses. No one has figured out how to make money on generative AI yet, but Google seems content spending more money to secure market share.

**RYAN WHITWAM**

Senior Technology Reporter

Ryan Whitwam is a senior technology reporter at Ars Technica, covering the ways Google, AI, and mobile technology continue to change the world. Over his 20-year career, he's written for Android Police, ExtremeTech, Wirecutter, NY Times, and more. He has reviewed more phones than most people will ever own. You can follow him on Bluesky, where you will see photos of his dozens of mechanical keyboards.