Neurorights

Neurorights are a proposed category of human rights protecting the brain and mental life against interference by neurotechnology, covering at minimum the privacy of neural data, freedom from unconsented modification of mental states, and control over one's own cognition. The concept was formulated in 2017 and has moved into law faster than almost any other proposal in technology governance — a constitutional amendment within four years, state statutes within seven, and an international recommendation within eight — largely before the technology it anticipates exists in deployable form.

§The proposed rights

Two overlapping formulations circulate.

Marcello Ienca and Roberto Andorno proposed four: cognitive liberty, the right to use or refuse neurotechnology; mental privacy, protection of neural information; mental integrity, freedom from unauthorised alteration of neural computation; and psychological continuity, protection of the sense of self against unconsented change.¹

The NeuroRights Foundation, founded by the Columbia neuroscientist Rafael Yuste out of the group that published a 2017 set of ethical priorities for neurotechnology in Nature, advocates five: mental privacy, personal identity, free will, fair access to mental augmentation, and protection from algorithmic bias.² The fourth restates the Human enhancement question as an entitlement rather than a liberty and connects the framework to Access and inequality; the fifth borrows from AI governance and fits awkwardly with the rest.

The two lists disagree on whether neurorights are new rights or applications of existing ones. Freedom of thought is already an absolute, non-derogable right under the International Covenant on Civil and Political Rights, and has almost never been litigated because until recently no technology could reach the thought itself.

What the rights are protecting against

Not one mechanism but three: reading neural signals without consent, writing to the brain without consent, and inferring mental content from data that is not obviously neural at all. The first is technically hard, the second is routine in medicine, and the third is the one already happening at scale.

§Why the proposal arrived when it did

Three developments converged. Implanted recording devices moved from laboratories to multi-year clinical use, described in Brain–computer interface and Electrocorticography interfaces. Consumer electroencephalography headsets became cheap enough to sell as meditation and focus trainers, generating neural data outside any medical-privacy regime. And decoding methods improved to the point where research groups could reconstruct approximate meaning from brain activity, a body of work covered in Neural decoding and Speech neuroprosthesis.

Stimulation supplies the second half of the concern. Deep brain stimulation is an established therapy in which a device alters neural activity continuously, and a minority of patients report changes in mood, impulsivity, or the sense that their actions are their own. Non-invasive neuromodulation devices are sold direct to consumers with no clinical oversight. Mental integrity is not a hypothetical right when a manufacturer can push a firmware update to an implanted stimulator.

§Law enacted so far

Chile amended its constitution in 2021, adding to the article on physical and mental integrity a clause directing that scientific and technological development respect life and physical and mental integrity, and that the law regulate the requirements for activity involving neural data and brain activity. It is the first constitutional protection of its kind. In 2023 the Chilean Supreme Court applied it, ruling against a consumer EEG manufacturer in a case brought by a former senator and ordering the deletion of his neural data — the only judicial decision anywhere resolving a neurorights claim on the merits.

US states have legislated through data-protection statutes rather than rights instruments. Colorado amended its privacy act in 2024 to bring biological and neural data within the definition of sensitive data requiring consent; California added neural data to the sensitive-information category of its consumer privacy act in the same year; further states followed in 2025. The approach is narrower than Chile's — it regulates processing, not interference — and it applies only to companies meeting the statutes' size thresholds.

International instruments are non-binding. The OECD issued a recommendation on responsible innovation in neurotechnology in 2019, the first intergovernmental instrument on the subject. UNESCO adopted a Recommendation on the Ethics of Neurotechnology in late 2025, following the model of its earlier AI ethics recommendation: member states report on implementation, and nothing obliges them to implement.

§Criticism

The main objection is that neurorights are redundant. Jan Christoph Bublitz and others argue that existing rights to privacy, bodily and mental integrity, and freedom of thought already cover the described harms, that creating new rights dilutes the human-rights framework, and that the proposals are drafted around imagined capabilities rather than actual ones.³ On this view the correct response to a company selling EEG headsets is enforcement of ordinary data-protection law, not a new constitutional category.

A second objection concerns definitions. "Neural data" has no settled meaning. If it covers only signals recorded directly from neural tissue, it excludes the inference channels that actually reveal mental states — eye movement, heart rate variability, keystroke dynamics, typing latency — and a company can escape the statute by using better sensors of the wrong kind. If it covers anything from which mental states may be inferred, it covers nearly all behavioural data and becomes unadministrable. Existing statutes have chosen the narrow definition.

A third is that the framing overstates what decoding can do. As Mental privacy sets out, the best non-invasive semantic decoding requires many hours of per-subject training data, a cooperative participant lying still in a scanner, and fails when the subject deliberately thinks about something else. Nothing resembling involuntary mind-reading exists. Advocates reply that rights are properly established before the capability arrives rather than after, which is a reasonable position but concedes the empirical point.

New rights or old ones enforced

Advocates argue that neural data is categorically different because it is generated involuntarily and can reveal states the subject has not chosen to express. Critics reply that the same is true of a consumer heart-rate monitor, and that treating the brain as special repeats the error diagnosed in genetic exceptionalism — see Genetic discrimination.

§Enforcement difficulty

Even where law exists, three features make it hard to apply.

Consent is the mechanism most statutes rely on, and it is doing work it cannot bear. A user agreeing to a headset's terms of service cannot evaluate what will be inferable from their data in five years, because the inference depends on models not yet trained.

Jurisdiction is the second problem. Neural data collected by a device sold in one country is processed by a service operated in another and may be sold to a broker in a third. Chile's constitutional protection binds conduct in Chile; the company in its Supreme Court case was foreign.

Medical and consumer devices are governed by different regimes with different assumptions. An implanted stimulator is a regulated medical device subject to clinical oversight; a wearable making similar claims about attention or mood is often a consumer product. The gap is where most current activity sits, including workplace fatigue-monitoring headsets used in transport and mining.

§Open problems

The distributive right — fair access to augmentation — has attracted the least analysis and would be the hardest to make justiciable. No court has any basis for ordering a state to provide cognitive enhancement, and it is not obvious what a violation would look like.

The relationship to therapy is unsettled. A right against unconsented alteration of mental states, read literally, is in tension with involuntary psychiatric treatment, which every legal system permits under conditions. Neurorights instruments have not addressed how the two fit. Nor have they addressed devices that write to memory rather than read from it: the hippocampal systems described in Memory prosthesis would, if they worked well, raise a question about the authenticity of recall that no privacy framework is shaped to answer.

And the timing question remains open. Constitutional protections written for capabilities that do not exist risk either freezing the wrong definition into law or accumulating as symbolic provisions that courts decline to apply. Chile's court did apply its provision, to a consumer headset whose recordings contain little of what the amendment was meant to protect. Whether the framework will be available and correctly shaped when devices of the kind described in Neuralink and Synchron research reach scale is not yet possible to judge.

§See also

• Mental privacy
• Brain–computer interface
• Neural decoding
• Deep brain stimulation
• Genetic discrimination
• Non-invasive neuromodulation
• Morphological freedom
• Access and inequality

§References