Google Arts & Culture Lab artist in residence, Gaël Hugo, explores what generative audio can offer for learning about cultural sites captured with Street View on Google Arts & Culture.

Select location on the map or from the list, explore the Street View, press the snapshot button and be guided by informative insights from Google AI. At any time, you can click on the ‘ask a question’ button which will generate three questions based on your Street View. This first experimentation is working with a small selection of partners and cultural sites, with more to be added in the future.

The experiment uses Gemini - Google’s Multimodal Large Language Model - to analyze Street View images, incorporating GPS coordinates and location names for context. Gemini then generates a script, which is sent to a Google AI Audio Model to generate the audio in real time.