Muse sleep band users weren’t told their brain data was training AI

In this article Muse’s maker used about 50,000 hours of users’ brainwave recordings to train AI without explicitly telling them Stripping names might not be enough to make brainwave data anonymous; researchers have identified individuals from EEG recordings Check privacy policies and ask about opt-out and data deletion Technology companies have spent years trying to get into your head for profit. Now they’re doing it literally. The maker of Muse, a headband that measures brain activity during sleep and meditation, used users’ brainwave data to train AI without explicitly telling them, an investigation has revealed. Companies sell these consumer sleep bands to help you better understand what your brain does when you sleep and meditate. However, according to an investigation in Science magazine, they’re picking up a little understanding of their own on the side. The year-long investigation found that Canadian company Interaxon, which makes the Muse sleep and meditation headband, was using the data that it collected from users’ brains to train AI. The company used about 50,000 hours of electroencephalography (EEG) recordings, which measure the brain’s electrical activity, to train a “brain foundation model.” AI models trained with large amounts of brainwave data from different people can spot patterns in brain activity. For example, a model might learn to recognize patterns associated with seizures. Researching such outcomes is laudable, but only if you first tell the people whose EEG data is being used for the training. Interaxon didn’t. This is your brain on an inadequate privacy policy When Science questioned Interaxon about the issue, its chief marketing officer Nadia Kumentas acknowledged that users should be explicitly informed about the AI training. The company updated its privacy policy, which now says de-identified data is used to train machine learning models. Collecting data through devices like this is tempting for companies because brainwave data is otherwise difficult to obtain and therefore incredibly valuable. No wonder, then, that it isn’t the only neuraltech company facing questions about its data practices. The publication reviewed 57 companies that collect brain data, and found that only six explicitly said they wouldn’t sell users’ brain data. More than two-thirds allow brain data to pass to a new owner if they are acquired. Eleven of them only had a privacy policy for website visitors, and four had no policy at all. To its credit, Interaxon removed names and email addresses from user EEG data before using it for training. However, those users’ data has already helped shape a model trained without explicitly telling them. Removing its influence after the fact is extremely difficult. But is such “de-identification” enough? An academic study conducted by the University of South-Eastern Norway this year found that EEG recordings and derived features can contain sensitive patient-specific information. It said that this can create privacy risks when that data is shared across clinical and research environments. It added: “Conventional anonymization methods are often insufficient for high-dimensional biomedical signals, since removing direct identifiers does not necessarily prevent re-identification, linkage, or inference risks.” Another study by researchers at Ben-Gurion University, Israel, found that analyzing resting-state EEG recordings could shed light on personality traits and cognitive abilities. Resting-state EEG identified users 88.4% of the time, according to a 2018 study by researchers at Korea University. And a University of Southern California study claimed that EEG data could reveal mental health status, neurological conditions, and even substance use history. These findings do not establish what can be inferred from Muse recordings specifically. Consumer headbands generally have fewer electrodes than clinical EEG systems. This isn’t the first review of the consumer neurotech industry to reveal some worrying trends. A 2024 Neurorights Foundation report found that 29 of 30 companies it surveyed appeared to have access to consumers’ neural data and had provided “no meaningful limitations to this access.” The law is still catching up to all this. The Illinois Biometric Information Privacy Act (BIPA) counts your fingerprint, voiceprint, or face geometry among other things as biometric identifiers, but EEG data doesn’t make the cut. However, Colorado and California do now treat neural data as sensitive. These concerns extend beyond brainwave data. In a Malwarebytes survey of newsletter readers , 90% worried about AI using their data without consent, while 92% were concerned about corporations using their personal data inappropriately. What Muse owners can do Do you use a sleep band or another neurotech device? Consider that your brain data might be used for more than you imagined. And if the company goes bust and/or gets acquired, it might pass to another owner. That’s exactly what happened with genetic testing company 23andMe. With this in mind, check the manufacturer’s privacy policy, and if you can’t find one, ask it directly. Ask whether your EEG data is used to train AI, whether you can opt out, and whether you can delete recordings already collected. If those options aren’t available, consider whether you’re comfortable continuing to use the device. From reporting threats to removing them. Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today . About the author Danny Bradbury has been a journalist specialising in technology since 1989 and a freelance writer since 1994. He covers a broad variety of technology issues for audiences ranging from consumers through to software developers and CIOs. He also ghostwrites articles for many C-suite business executives in the technology sector. He hails from the UK but now lives in Western Canada.