Can AI learn to smell - and turn a text prompt into a real fragrance? Osmo is connecting chemical measurements with the way people describe odors. This short explainer follows the process from human scent descriptions to AI-assisted formulas and physical samples, then explores a surprising application: investigating counterfeit sneakers.
The important distinction: software does not sniff the world on its own. Instruments measure chemical composition, people supply descriptions of perception, and perfumers help refine fragrances. Osmo's claim of "human-level olfactory intelligence" is presented here as the company's claim, not an independently reproduced universal benchmark.
We also look at why context matters. As researcher Michael Schmuker explains in MIT Technology Review's reporting, some odor molecules can be shared by very different things. Chemistry is part of the story; human interpretation adds more.
Sources and further reading:
MIT Technology Review, Amit Katwala, 18 November 2026:
https://www.technologyreview.com/2026/11/18/1148816/smell-startup-osmo-ai-nose-good-as-humans/
Osmo official website and fragrance workflow:
https://www.osmo.ai/
Osmo founder's explanation of AI and perfumers:
https://www.osmo.ai/blog/introducing-generation-fragrance-design-for-all
Visuals combine credited Osmo public-site photographs and official promotional demonstrations with original illustrations and animated explanatory diagrams. The sneaker artwork is AI-generated; diagrams are illustrative, not measurements, actual model outputs or footage of a product test. English narration is synthetic; music is an original local composition. No fragrance service, sneaker detector or scientific benchmark was personally tested for this film.
What would you ask an AI perfumer to create?
More visual technology stories: https://www.youtube.com/@ABV_Creative
TikTok: https://www.tiktok.com/@lumierenoire.films
#Osmo #ArtificialIntelligence #ScienceExplained
YouTube / ABVX
