Tech
This smell startup says its AI nose is as good as a human’s
Publication: MIT Technology Review
Author: Amit Katwala
Source published: 18/11/2026, 10:20:30 UTC
Briefing created: 18/11/2026, 11:29:48 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 18/11/2026, 10:35:24 UTC.
Briefings
Osmo AI, a startup spun out of Google Brain, claims its foundation model for smell achieves human-level olfactory intelligence. The company claims it has achieved this by analyzing over 10 million human labels and 800 million scent tokens collected via chemical sensors. While the startup claims its technology outperforms in-house perfumers and could be used for threat detection, critics note the heavy reliance on human sensory panels and lack of independent verification. The research is based on data collected in a lab near Newark Airport, New Jersey, with limited public context regarding the specific deployment scenarios.
Key points
- Osmo AI claims its foundation model for smell achieves human-level olfactory intelligence, outperforming in-house perfumers at a major consumer goods company.
- The company claims it has analyzed over 10 million human labels and 800 million scent tokens to create its model.
Context
Osmo AI is a startup spun out of Google Brain's former AI research lab, founded by Alex Wiltschko. The company claims its technology is already being used in the fragrance industry to help brands design and manufacture new scents, with a grant from the Gates Foundation supporting insect repellent research.
Limitations
Public context is limited to the lab near Newark Airport, New Jersey, where the team of 30 human sniffers operates. The company's specific deployment scenarios for threat detection and disease diagnosis are not detailed in the provided evidence, and independent verification from external experts remains unconfirmed.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 18/11/2026, 11:29:48 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeibf6vfijjhdyq4yxzuihtpsuaoqytfnwzmct2jdlnruzfbmtmithq