Related video: Ornn's Kush Bavaria on building a financial market for GPU compute, @20VC
Trending on MIT Technology ReviewOsmo's AI smell model aims to match human olfactory ability
MIT Technology Review AI · · 25-second read
Top comments from YouTube.
This story is from November 18, 2026. For the latest, see today's news.
November 18, 2026 · AI Startups
Osmo, an AI startup, has built what it claims is the first foundation model for smell by training on data from about 30 professional human sniffers who document odors alongside chemical analysis. The company has created a large dataset of olfactory descriptors—words like "nutty" and "peachy"—paired with molecular composition to teach AI to recognize and predict smells.
The model is already deployed in fragrance development. Osmo is now exploring applications in medical diagnostics and security threat detection, though questions remain about whether AI can reliably replicate human sensory judgment in those higher-stakes domains.
This is an AI-written summary. The original reporting belongs to MIT Technology Review AI. How we write these · Spot a mistake?
Get the AI brief every morning.
The five AI stories that matter each morning: startups, money, jobs, tech and what's coming next. Free, and one click to unsubscribe. See our Privacy Policy.