The announcement comes amid scrutiny of Google's Gemini AI model for creating historically inaccurate images of people. The new health model will initially be launched as experimental Labs features, allowing Google to gather feedback before wider rollout. It is unclear when the model will be released and which products it will be featured in. There are also questions about whether the model will run on-device or require an internet connection. More information is expected to be revealed at Google I/O, the company's annual developers conference.
Key takeaways:
- Google is developing a Personal Health Large Language Model based on its Gemini models to analyze Fitbit data and provide wellness advice and recommendations.
- The new model will power Google's upcoming experimental Fitbit Labs program and other AI features in the future.
- The model is trained using de-identified health signals from research case studies, validated with accredited coaches and wellness experts.
- There are concerns about the reliability of the information provided by generative AI models and chatbots, as they can struggle with issues like hallucination and bias.