Flo Health builds safer AI models for health app with Databricks
Flo Health, maker of a women’s health app with 81 million monthly users, now uses custom-trained generative AI models—rather than generic ones—built on Databricks’ platform. By defining its own standards for medical safety and usefulness, then training models to meet them, Flo aims for better, safer answers. This move is expected to cut AI operation costs significantly while giving Flo tighter control over quality.
- Flo shifted to open, fine-tuned generative AI models
- Databricks AI Runtime enabled quick, scalable training
- Evaluation focuses on both medical safety and usefulness
- Flo’s system uses AI judges to enforce quality standards
- New approach cuts AI costs by up to tenfold
Sources covering this
In this story
More in AI
Meta launches Muse Wonder, its most advanced AI yet
Meta just launched Muse Wonder and Muse Wonder Full, its most advanced AI models to date—and its first so-called 'frontier models' meant…
Microsoft president urges stronger AI regulation
Microsoft president Brad Smith has called for immediate, stronger regulations on artificial intelligence, arguing that AI is less…
Epic Games to use AI for code and art in Disney shooter
Epic Games is reportedly adopting an 'AI first' approach for its next collaborative shooter with Disney, using generative AI to create…
Google highlights AI-powered disease research collaborations
Google is working with partners to use AI for medical research, focusing on making diseases easier to detect and treat.