Google founder Sergey Brin cautions against using Gemini Live while driving

Google co-founder Sergey Brin has cautioned users against relying on Gemini Live while driving, stating that the current public version of the AI tool is not yet suitable for situations requiring full attention, according to a report by Business Insider.

Speaking during a panel discussion at Stanford University, Brin said the version of Gemini Live currently available to the public runs on an outdated model and should not be used in high-risk scenarios such as driving. He informed the audience that while he personally tests Gemini Live during his commute, this practice should not be interpreted as guidance for users.

Brin said he engages in what is commonly referred to in the technology industry as dogfooding, where developers use their own products to understand real-world performance. During these tests, he said he poses complex questions to the AI, including queries related to estimating the cost and power requirements for building data centres, as reported by Business Insider.

However, Brin stressed that users should avoid doing the same, stating that a significantly improved version of Gemini Live is expected soon. He said the publicly available model was an older iteration and described it as ancient compared with what Google is currently testing internally.

According to Brin, there is a substantial capability gap between the internal version of Gemini Live and the one released to users. He informed that the internal model handles complex questions far more effectively and that a major public upgrade is expected within weeks. The upcoming release is likely to be powered by Gemini 3, Google’s next-generation AI model.

The warning comes amid growing use of voice-based AI tools in daily activities, including commuting. Brin’s remarks indicate that Google is conscious of the distraction risks posed by conversational AI and is urging users to wait until the technology matures further before deploying it in safety-critical environments.

Gemini Live is designed as a conversational AI system similar to Google Assistant but with more advanced capabilities. It allows users to speak naturally, interrupt responses, and ask follow-up questions without restarting conversations. The tool also offers multiple voice options and adjustable response tones.

The platform is positioned for everyday use cases such as interview preparation, skill assessment and conversational practice, with the aim of making interactions with AI feel more fluid and human-like.