Mental privacy · Grey Matter

Mental privacy is the claim that what happens in a person's mind, and the brain activity that carries it, should be protected from being read, recorded or inferred without their consent, and it is the first ethical question that any technology for reading the brain raises.

Mental privacy. Mental privacy is the claim that what happens in a person's mind, and the brain activity that carries it, should be protected from being read, recorded or inferred without their consent, and it is the first ethical question that any technology for reading the brain raises.

Until recently the claim protected itself: thoughts were inaccessible, so privacy law could focus on what people said, wrote or did. Devices that record the brain change that. Even crude signals reveal something (attention, drowsiness, recognition of a familiar face, the response to an image), and decoders trained on richer recordings have reconstructed the gist of speech a person heard or imagined from fMRI, and some inner speech from implants in motor cortex. The concern is less that minds can be read in full today than that brain data, once collected, can be analysed again later with better decoders, combined with other data, and used for purposes the person never agreed to.

It is proposed as a right of its own. Ethicists have argued that existing privacy and data law does not cover the mind well, and proposed rights to mental privacy, cognitive liberty, mental integrity and psychological continuity; see neurorights.

Current decoders need cooperation. The fMRI decoder of 2023 had to be trained on many hours of the same person's data and failed when the person resisted, which is a technical safeguard today and a weak legal one, since it may not last.

Privacy covers inference as well as content. A recording that reveals a disease, a mood or a preference exposes the person even when no word is decoded, so laws that protect neural data aim at the data itself.

Mental privacy has to be protected before it can be violated at scale.

Once neural data is collected and stored, what it reveals grows with every improvement in decoding.

Questions: Could a speech implant pick up words its user never meant to say aloud? A study of four people with speech implants in motor cortex found that imagined sentences could be decoded in real time, and that some free-form inner speech could also be decoded while participants recalled sequences or counted silently. The same study demonstrated ways to prevent it, among them keeping decoding locked until the user imagines a chosen keyword. It is direct evidence that a device built to restore communication can capture more than the user means to share, which turns mental privacy into a design requirement for speech implants. Can fMRI decode what a person is thinking without their cooperation? In 2023 a decoder trained on about 16 hours of a person's fMRI recordings reconstructed the gist of stories they listened to or imagined, and of silent films they watched. It worked only for the person it was trained on and only with their cooperation: decoders trained on other people failed, and participants could defeat it by thinking of something else. The authors presented those limits as a current safeguard and called for policies to protect mental privacy before decoders become less dependent on cooperation. Can functional ultrasound decode what someone is about to do, and what would a wearable version mean for privacy? In monkeys recorded through an opening in the skull over the posterior parietal cortex, functional ultrasound predicted from single trials the direction and the effector (eye or hand) of a movement the animal was planning, at about 100 µm resolution. In people it has so far worked only where bone was removed or replaced, as through a transparent skull implant after an injury. A version that worked through intact bone would read intentions in depth without surgery, which is why it is watched as a likely future case for mental privacy rules written for neural data. Could microwaves passed through the head read what a person is thinking? With the physics as it is known, no. Microwave tomography maps how much water and salt tissue holds, at grids of a few millimetres and wavelengths of centimetres inside the head, which is enough to find a pool of blood after a stroke and far too coarse and slow to follow neural activity, whose effect on those properties is tiny. A worry about mental privacy is better spent on the methods that do read activity, such as EEG headbands, fMRI decoders and implants, where the data already exist. Does a wider data stream out of the brain expose more of the mind? More channels and richer features let a decoder find more than the one thing it was built for. Speech implants in motor cortex, designed to decode attempted speech, also picked up some inner speech during tasks such as counting, and the more of the raw signal leaves the head the more such secondary readings become possible later. Compressing data on the implant to only what one decoder needs is an engineering choice that also limits exposure, and ethicists have proposed safeguards of this kind, processing neural data where it is collected and sharing only what a service needs. Does the physics of the skull protect mental privacy? For now, partly. Scattering of light, attenuation of ultrasound in bone and the weakness of the brain's fields outside the head mean that no device can read detailed brain activity from a distance or without the person's knowledge, and even fMRI decoding of meaning needs hours of cooperative training inside a scanner. That protection is a fact of current engineering and can erode, because wearable sensors improve and decoders extract more from blurred signals every year. Ethicists and lawmakers therefore treat physics as a temporary safeguard and write protections for neural data before the technology removes it. Which rights have been proposed to protect the mind, and where does mental privacy sit among them? Ienca and Andorno proposed four rights in 2017: cognitive liberty, mental privacy, mental integrity and psychological continuity. The group around Rafael Yuste, which later founded the NeuroRights Foundation, proposes five: mental privacy, personal identity, free will, fair access to mental augmentation and protection from algorithmic bias. Mental privacy is the one shared by every list and the one closest to law, since it can build on existing data protection, while the rights that protect identity and free will address devices that write to the brain and are harder to define. Why is neural data treated as more sensitive than other personal data? A recording of the brain can reveal things its owner has not chosen to express and may not know, such as recognition, attention, mood or signs of a neurological condition, and it is produced without a deliberate act the way speech or typing is. What it reveals also grows after collection, because a stored recording can be analysed again with better decoders years later. For these reasons ethicists proposed treating it like organs, protected by default and shared only by explicit choice, and Colorado and California classed it as sensitive data in 2024.