How it works
Users can type, speak, or upload photos (e.g. a maths problem) to get an explanation. The app can talk back by reading answers aloud. It can be used as a coach to solve a problem or learn about something. However, generative AI chatbots can confidently present false information as fact.
A user can ask Gemini to research and summarise a topic or a long article. They can use ‘Gems’ to chat about specific topics. Gemini Live (real-time voice conversation) is restricted to users aged 18+. On Android, Gemini can perform AI tasks like drafting a message. It can search YouTube for a specific video or find a location on Google Maps. Keeping prompts short and only creating images when necessary helps reduce the environmental impact of AI models.
Parental controls or safety settings available
A user can report harmful content on the app. Gemini has no built-in parental controls. A user can prevent Google from using their chat history to train its AI models, but this also deletes their chat history. A paid-for business account is needed to retain chat history without allowing its use for training.
Anyone can access a guest browser version of Gemini without logging in. This includes standard filters for explicit content. Children with a Google account aged under 13 cannot use the Gemini app. Children aged 13-17 can access the Gemini app (unless they have a ‘supervised’ Google account). Teenage accounts get automated teen safety filters that are more restrictive.