Osmo trains AI smell model and applies it to fragrance design
Osmo is training an AI model using odor descriptions paired with chemical analyses, according to MIT Technology Review. Founder Alex Wiltschko claims human-level smell capabilities, though the report notes skepticism about the technology.
AI Hubs United Editorial ·
AI startup Osmo is developing a foundation model for smell and is already helping fragrance brands create scents, according to MIT Technology Review. Researchers in a New Jersey laboratory are recording descriptions of odors and matching those observations with chemical analyses to train the system. Founder Alex Wiltschko says the aim is to give computers a sense of smell and claims the model has reached human-level olfactory intelligence.
The work connects human assessments of scent with chemical information, with fragrance design providing a practical application for the model. Osmo also hopes to apply its technology to identifying counterfeit products and diagnosing diseases. Those uses are ambitions rather than reported deployments. MIT Technology Review notes that not everyone is persuaded the technology can deliver, an important distinction between the founder’s capability claim, its reported use by fragrance brands, and the broader applications the company wants to pursue.
AI Hubs United original short summary of the source linked above. It is not the original article; read the full story at the source.
