Google could be secretly training Gemini to focus on user 'Projects'

What you need to know

• Google is reportedly working on bringing "Projects" to the Gemini app, a new feature that users can lean on to focus the AI on a singular topic.
• From the leaked code, it seems users can upload "up to 10" files that the AI can leverage, alongside your set of specific instructions for the project.
• Recent leaks claim that Google is looking to add NotebookLM notebook importing to the Gemini app.

Google is reportedly developing a new feature for the Gemini app that could bring the AI forward in a big way for your larger "projects."

A new feature in the Google app's latest version was spotted with ties to the Gemini app called "Projects" by tipster AssembleDebug and Android Authority. The tipster discovered that Google is looking to place this "Projects" feature in the side panel of the Gemini app. From there, users can tap "New Project" and begin working, but Google's development gives a small description of the feature.

According to the post, Google's code states users can "Add up to 10 files that Gemini can reference throughout your project."

The publication likens this supposed Gemini feature to OpenAI's "Projects in ChatGPT." The company's support page for this feature states this feature is all about context, as users can upload files, instructions, and more for the AI in that project session to help keep it "on topic." Speculation claims that Google is moving in this direction, too, which is furthered by the development's little tid-bit: "projects keep your work and ideas organized. You can add context and tell Gemini how to help."

Google is seemingly continuing down this path of a more personalized or "focused" AI, kind of like NotebookLM (in a way).

At the moment, the tipster was unable to surface more of Gemini's Projects, specifically, how this would work. Right now, they were able to get as far as the "Give your project a name" screen, right before "create project" becomes accessible. It's probably safe to say Google will have a more expansive screen after this, one that gives users a chance to upload their files and enter in their instructions, per the leaked descriptors.

A more personal, focused AI

Funnily enough, it seems that Google is working on bringing NotebookLM to Gemini in a future update; at least, that's what the rumors suggest. It seems that the company will expand its "Attachments" menu in the Gemini app to include the option to import your NotebookLM notebooks. This importing feature would also include sources, if this actually progresses to a stable launch.

Google is really trying to blur the lines between Gemini and its focused NotebookLM model to help users research, understand, and sift through heavy information and research. Each side has its strengths, which is also something Google may want to capitalize on by merging them and showing users their varying uses all in one place.

Other updates for Gemini include a more personalized experience, which is tailored to the user and their interests/needs, based on their Search behavior. This was announced as an experimental feature earlier this year.