Interaxon, the Canadian company behind the Muse EEG headband, has trained an AI model of the brain on about 50,000 hours of electroencephalography (EEG) data from its users, the company told Science in an investigation published on 4 November. Its privacy policy did not explicitly tell users their data might be used this way; after Science pointed this out, the company updated the policy.
Muse here is the sleep and meditation headband, not Meta’s Muse AI agent. Science reviewed the privacy policies of 57 consumer companies that collect EEG data. By our arithmetic, 42 had a policy that went beyond website visitors, and Science found only six that explicitly said they would not sell users’ brain data, while another 17 said in general terms that they do not sell personal information.
Business Pill · Consent
An evergreen Business Engineer explainer on consent: what it means for a company to have permission to use personal data for a given purpose. It explains the general idea only and does not discuss Interaxon, Muse or any company in this piece.
The key insight: Interaxon, maker of the Muse EEG headband, told Science it has trained a brain foundation model on about 50,000 hours of user EEG data, and updated its privacy policy after Science pointed out it did not explicitly disclose this. Science reviewed 57 consumer EEG companies; by our arithmetic 42 had a broader policy, and Science counted six that explicitly rule out selling brain data.
Interaxon is building what the industry calls a “brain foundation model”, Science reports, to sharpen its picture of how the brain works, especially during sleep and times of focus. The data are stripped of personal details such as names and emails, and the company plans to keep expanding the training set. “We aim to use as much of that eligible corpus as possible,” Chief Marketing Officer Nadia Kumentas told Science.
Science reports that the policy did not explicitly mention training AI brain models until it raised the point. “We agree users should be explicitly informed,” Kumentas said. “Our brain data is the most valuable part of the company,” Kumentas also told Science.
Science reviewed the privacy policies of 57 consumer companies that collect EEG data. The 42 broader policies are our arithmetic from Science’s figures; the six explicit no-sale policies are Science’s count.
What the Policy Says Now
The privacy section on Muse’s legal page, read on 9 November, says the company uses session data for purposes including “the improvement of our own algorithms and machine-learning models, including foundation models, that power Muse features such as sleep staging and brain-health metrics”. It adds: “Data used for this purpose is first de-identified.” The page does not show a date for that change.
The Wider Industry Review
Of the 57 companies Science reviewed, four had no privacy policy at all and 11 had a policy covering only website visitor data. Among those with a broader policy, about two-thirds explicitly mentioned collecting brain data. Science found that nine consumer neurotechnology and AI software companies have publicly acknowledged building a brain foundation model.
Initially none of the companies building such a model disclosed it explicitly in their privacy policies, Science reports. Emotiv told Science it trains its “large brain model” in part on data from users of its higher-end devices, and updated its policy after Science’s inquiries to disclose de-identified training subject to opt-in consent.
As we read Science’s account, the disclosure gap was in the policy text, not in the training itself: Interaxon says the data were stripped of names and emails before use.
As we read the policy page now, foundation-model training is listed among the purposes for session data, alongside research and product improvement, with de-identification as the stated condition.
As we read the industry review, Interaxon is one of nine firms Science says have publicly acknowledged building a brain foundation model, and Emotiv made a similar policy change with opt-in consent.
Nadia Kumentas, chief marketing officer, Interaxon, to Science
“We agree users should be explicitly informed.”
Three Implications
FOR MUSE USERS Interaxon’s policy now names foundation models among the uses of de-identified session data.
FOR THE INDUSTRY Science found that only six of the 57 companies it reviewed explicitly say they will not sell users’ brain data.
FOR REGULATORS Science reports that Colorado and California have amended state privacy laws to treat neural data as sensitive personal information.
The framework argues: “Each moat is built by a different data pipeline.”
As we read it, Interaxon’s 50,000 hours of user EEG, the company’s figure, is that kind of pipeline, and the policy change shows the consent terms are part of it. The framework predates this report and does not discuss Interaxon or neurotechnology.
What Is Not Established
Science does not report how many Muse users’ data went into the 50,000 hours, and Interaxon’s figure is the company’s own. The policy page we read does not date the change or say whether training requires opt-in consent. The count of 42 broader policies is our arithmetic from Science’s figures, and the six no-sale policies are Science’s count. We did not contact Interaxon.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Interaxon says it has trained a brain model on about 50,000 hours of Muse users’ EEG data and now discloses foundation-model training in its privacy policy, after Science flagged that the old policy did not. Science found that six of the consumer EEG companies it reviewed explicitly rule out selling brain data.
A note on sourcing. Interaxon’s figures and quotes, and the industry review, are from Science’s investigation of 4 November 2026, read in full. The policy wording is from the privacy section of Interaxon’s Muse legal page, read on 9 November 2026. The 42 broader policies are our arithmetic. We did not contact Interaxon or Science.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
