Gemini Notebook now reads aloud to students

Google is adding voice conversations, lecture recording, interactive learning paths, new quizzes, and short video summaries to Gemini Notebook.

A student can now open class notes on a phone, ask for a spoken explanation of Newton’s first law, interrupt when something is unclear, and continue the conversation with a more specific question. Gemini Notebook is no longer limited to organizing and summarizing documents. Google wants to turn it into a study assistant users can talk to.

The update arrives two months after NotebookLM was renamed. Since July 2026, the service has been called Gemini Notebook. It remains a standalone app, but its content can also be accessed through Gemini. Google has also announced plans to bring it to AI Mode in Search.

The latest release focuses on learning. It combines a redesigned interface, real-time voice conversations, an audio recorder, interactive learning overviews, additional quiz formats, and short educational videos. Not every feature will become available to every account at the same time.

Voice conversation is the most visible change. From the Android or iOS app, users can hold real-time discussions with a notebook in nearly 100 languages. They can ask questions by speaking, listen to the responses, and interrupt the assistant before it finishes explaining.

The feature brings the service closer to an oral tutoring session. A student can request a definition, point out that a step is unclear, ask for a simpler example, or return to a term mentioned a few seconds earlier. The conversation continues to use the documents collected in the notebook as its reference material.

Gemini Notebook does not answer solely from the model’s general knowledge. It uses the sources selected by the user and provides references that make verification easier. Students can import course materials, articles, webpages, ebooks, public videos, or their own notes before starting a conversation.

This deliberate restriction of context remains one of the product’s main strengths. A general-purpose assistant may mix the requested information with outside knowledge or respond from an overly broad interpretation. Gemini Notebook searches the notebook’s own collection first.

The approach does not guarantee an accurate answer. The model may connect two passages incorrectly, oversimplify a concept, or produce an explanation that does not match the expected academic level. The references are there to help users return to the original document and verify the interpretation.

Voice conversations are beginning to roll out to Google AI Ultra subscribers aged 18 and older. Google plans to extend the feature to Google AI Pro subscribers and other accounts later. The phrase “coming soon” therefore means that some users will not immediately see the option in their app.

The phone is also becoming a collection tool. Beginning the week after the announcement, a new recorder is expected to capture lectures, personal reflections, or ideas spoken while on the move. The recording then joins the notebook’s other resources.

The mobile app already supported photographs of handwritten notes, worksheets, and other study materials. Adding audio removes the need to use a separate app before importing a file. The resulting notes can be viewed, edited, cited, and queried from the same workspace.

Google says these notes can be managed on the web and mobile, including by eligible users under 18, as long as English is selected as the output language. The announcement does not say when this language restriction will be removed.

Recording a speaker also does not remove the need to follow an institution’s rules or obtain any required consent. The presence of a recording button does not mean a class, meeting, or another person’s voice can be stored and shared without permission.

Once the sources have been collected, students can create an “interactive learning overview” from the Reports section. It combines several outputs into a single path, such as a summary, infographic, flashcards, and quiz.

The idea is to avoid generating each resource separately. A notebook can prepare a general overview of a topic, offer a visual explanation, test the student’s knowledge, and help identify concepts that still need work. Users retain control over which materials they want to create.

Quizzes are gaining additional response types. Short answers, multiple-select questions, and fill-in-the-blank exercises will join the existing formats. These options cover different learning needs, from recognizing a concept to recalling it without a prompt.

After completing a quiz, students can ask the notebook chat about their performance. The notebook can identify missed questions, return to the relevant sources, and suggest which sections to review. Flashcards and quizzes remain editable, so users can correct a question or add a new one.

This customization prevents the first generated version from being treated as final. An ambiguous, overly simple, or off-topic question can be revised manually. Teachers and students retain an editing role instead of receiving a fully locked exercise.

The educational quality will still depend on the material provided. A quiz created from an incomplete document will reproduce its gaps. If two sources conflict, the model may also build a question around disputed information without properly reflecting its significance.

Performance tracking should not be confused with formal academic assessment. Google does not present Gemini Notebook as a certified tool for assigning grades, supervising exams, or measuring a class’s long-term progress. The feature is primarily intended to guide individual study sessions.

Interactive learning overviews and the new quiz formats will roll out to all users over the weeks following the announcement. The gradual release may temporarily leave two free accounts in the same country with different feature sets.

Google is also adding Short Video Overviews. These roughly one-minute clips combine narration with educational animations designed to explain a formula, diagram, or scientific concept. Support for more than 80 languages is promised.

The videos