Ethical, Legal, and Societal Considerations

Before debating whether enhancement is right or wrong, define the target: which capacity, changed by what mechanism, for which person, in which setting, for how long, and compared with what alternative?

The familiar division between therapy and enhancement is useful, but incomplete. Restoring speech through a brain–computer interface can be treatment, communication support, and a new human–machine capability at once. Correcting severe sleep deprivation may improve cognition more than an experimental device. A tool can also be therapeutic for one person, optional augmentation for another, and coercive surveillance in a third setting.

The goal is not a population that performs one task faster on command. It is people with greater power to understand, choose, invent, relate, and direct their own lives.

A person with exceptional reasoning, memory, imagination, insight, or accumulated expertise may be able to see possibilities, solve problems, or guide others in ways that cannot be reproduced simply by buying a tool, reading a summary, or assembling more people.

Tools can multiply human thought, but they do not erase the importance of the thinker who knows which question matters, detects a false assumption, combines distant ideas, anticipates consequences, or develops judgement over decades. When such a mind is damaged by preventable illness, chronic stress, violence, intoxicants, pollution, sleep deprivation, exclusion, or forced conformity, the loss is personal and social. Protecting cognitive health, quiet, autonomy, education, and meaningful work is therefore a legitimate public priority.

A valid higher IQ score can matter greatly. Within the abilities a well-designed test actually measures, a higher score is evidence of stronger performance—not a decorative label. In fields where rapid learning, abstraction, complex reasoning, or the integration of many variables is decisive, that capacity can make a person’s presence and contribution unusually important. Such ability may reflect inherited potential, years of education, a lifetime of disciplined learning, or all of these together. It deserves recognition, protection, demanding opportunities, and respect. A score alone does not guarantee wisdom or moral judgement, but neither should fear of hierarchy be used to deny real cognitive differences or erase the value of an exceptional mind.

The ethical question is therefore larger than “Should enhancement be allowed?” It is also: Are we already suppressing human potential through poor education, preventable disease, intoxicating environments, fear, coercion, or the destruction of attention? The safest and most widely shared cognitive gains may come from protecting brains and learning conditions before adding invasive technology.

The first protection against hype is to label the status of a claim accurately. “Researchers decoded attempted speech under tightly controlled conditions” does not mean a headset can read unrestricted private thoughts. “A device is being studied” does not mean it is authorised for sale. “FDA-cleared for one clinical indication” does not mean proven to raise intelligence in healthy users.

• Was the outcome an everyday capability or only a narrow test result?
• Was the comparison blinded, sham-controlled, and adequately powered?
• Were harms, dropouts, sleep, mood, and later performance measured?
• How large was the improvement—not merely whether it was “significant”?
• Did the effect replicate independently and persist after training or stimulation?
• Was the product used exactly as authorised, or is the claim an extrapolation?

International bioethics principles place human dignity, autonomy, consent, benefit and harm, justice, privacy, and non-discrimination at the centre of biomedical decisions.² The 2024 revision of the World Medical Association’s Declaration of Helsinki likewise treats freely given informed consent and added protection for people in situations of vulnerability as central to medical research ethics.³ A signature is evidence that a form was signed; it is not, by itself, proof that these conditions existed.

Capacity is decision-specific, not a label on a whole person

A person may need help understanding a complex implant while retaining strong ability to decide many other matters. Good practice offers plain language, interpretation, communication aids, extra time, trusted support chosen by the person, and checks for understanding. It does not confuse disability, unusual communication, distress, age, or disagreement with incapacity.

Children and future persons require stronger protection

Children can participate in decisions according to their developing abilities, but adults still carry duties to protect their open future. The more irreversible, uncertain, identity-shaping, or non-medically necessary an intervention is, the weaker the case for making it before the person can decide. Heritable genome editing creates an additional problem: the people affected across generations cannot consent at all.

BCI progress is real—and narrower than the headlines

A 2024 U.S. Government Accountability Office technology assessment reported that implanted BCIs in clinical trials had enabled communication and control functions while identifying surgery, device durability, cybersecurity, long-term support, access, and governance as continuing challenges.¹⁰ In March 2026, China’s medical-products regulator announced marketing authorisation for an implantable BCI system with a narrowly defined hand-movement compensation indication.¹¹ That milestone does not turn investigational speech decoding, consumer headsets, or cognitive enhancement into the same product class.

CRISPR should be described through what has actually been achieved. In December 2023, the FDA approved Casgevy, the first FDA-authorised therapy to use CRISPR/Cas9 genome editing. It edits a patient’s own blood-forming stem cells for sickle cell disease; it does not edit embryos, pass changes to descendants, or increase intelligence.¹³

Why intelligence is not currently an editable target

Intelligence is a complex, polygenic family of traits shaped by many genetic variants and by development and environment. “Polygenic” means that many genes contribute to variation in a characteristic. One large genome-wide study identified 205 associated loci, yet its polygenic scores explained at most 5.2% of intelligence variance in independent samples.¹⁴ That predictive limit does not mean biology is unimportant; it means that statistical associations do not provide a safe set of causal switches for producing a chosen cognitive outcome. A variant may affect several traits; effects may depend on other variants and environments; population-based predictions may transfer poorly across ancestry groups; and editing embryos can introduce unintended on-target, off-target, mosaic, or chromosomal changes.¹⁵

Embryo polygenic scoring and embryo genome editing are also different. A score estimates relative probabilities from inherited variants among available embryos; it does not add a trait, guarantee an outcome, or demonstrate that editing would be safe. As of this article’s evidence cut-off, there was no clinically validated embryo-CRISPR method for increasing IQ, and accepted international guidance did not support reproductive use for cognitive enhancement.¹⁵

The international commission convened by the U.S. National Academies and the Royal Society concluded in 2020 that heritable human genome editing was not ready for clinical use; even a hypothetical initial pathway was confined to tightly defined serious monogenic disease, not polygenic enhancement.¹⁵ WHO’s 2021 governance framework covers somatic, germline, and heritable editing through registries, coordination, public engagement, and action against unsafe or unethical work.¹⁶ It is guidance, not a worldwide licence or global criminal code.

The deepest risk is not that some people will freely pursue greater capability. It is that institutions will define one preferred cognitive style, measure it continuously, and make safety, income, education, or belonging conditional on biological compliance.

UNESCO’s 2025 Recommendation says workplace neurotechnology should be actively voluntary and refusal should not bring adverse consequences. It also urges strong limits in education, including no academic penalty for refusal and no use for performance evaluation. For healthy people under 18, it advises against non-therapeutic use for optimisation.¹

Autonomy includes the freedom to become more capable

An autonomy-based case for enhancement deserves respect: adults may have strong reasons to improve attention, communication, memory, or other capabilities. But this is a qualified freedom, not a general legal right to any product. Safety, evidence, law, effects on others, and fair access still matter. Most importantly, freedom to enhance requires an equally protected freedom not to enhance.

Dual use follows the data and the power—not only the hardware

A system developed for communication or rehabilitation may later be repurposed for workplace monitoring, unconsented assessment, insurance discrimination, military optimisation, interrogation, advertising, or behavioural influence. The OECD’s Recommendation on Responsible Innovation in Neurotechnology asks governments and innovators to build safety, inclusion, public deliberation, brain-data protection, stewardship, and monitoring of misuse into development from the start.¹⁸

The ethical problem is not only what a system can infer today. It is what stored data may reveal tomorrow, who can combine them with other records, and whether the person can challenge an inference that affects employment, insurance, education, healthcare, credit, or liberty.

Under the EU General Data Protection Regulation, neural data are not a standalone statutory category. If they relate to an identified or identifiable person, they are personal data. They receive Article 9 special-category protection only when the statutory definition is met—for example, as health data, genetic data, or biometric data processed for unique identification.¹⁹ UNESCO’s neurotechnology framework goes further as policy guidance by recommending strong protection for neural data and mental-state inferences, including purpose limitation, minimisation, security, and meaningful rights of control.

The United States has a sectoral privacy landscape

In the United States, HIPAA’s Privacy Rule protects protected health information held or transmitted by covered entities and their business associates; it does not automatically cover all health information or every consumer wellness or neurotechnology company. HHS provides a tool for developers to identify whether HIPAA, the FTC Act, the FTC Health Breach Notification Rule, medical-device law, or other federal rules may apply.²⁰ The FTC’s amended Health Breach Notification Rule applies to qualifying vendors of personal health records and related entities outside HIPAA and requires notice after a covered breach, including unauthorised acquisition of unsecured PHR-identifiable health information.²¹

State law is evolving. Colorado amended its privacy law in 2024 to address biological data, including neural data, and California amended its consumer privacy law to include neural data within sensitive personal information.²²²³ These are important developments, not proof that the United States now has one comprehensive national neuroprivacy code.

The first equity obligation is to stop wasting existing human potential. Many people still lack quality education, diagnosis, hearing and vision support, rehabilitation, assistive technology, digital access, clean environments, adequate sleep, nutrition, safety, or freedom from substances and conditions that damage cognition.

The WHO and UNICEF global report on assistive technology found major access gaps and frames appropriate assistive products, services, trained personnel, and policy as essential to participation and inclusion.²⁴ A society discussing premium neural augmentation while people cannot obtain a hearing aid, communication device, or evidence-based treatment has not solved the basic allocation problem.

Ask who is represented in both evidence and design

A system trained or tested on a narrow population may perform differently across bodies, brains, languages, skin and hair characteristics, disabilities, ages, cultural practices, and clinical conditions. Exclusion from trials creates uncertainty; inclusion without shared power can merely redistribute risk. Responsible programmes pay community partners, make materials accessible, publish subgroup performance, and allow people affected by the technology to influence outcomes and stopping rules.

Justice should raise the floor of cognitive health and opportunity. It should not lower the ceiling of human learning, originality, or excellence.

Regulators rarely govern “cognitive enhancement” as one category. Legal responsibility depends on what the maker claims, how the product works, who uses it, whether it is medical, what data it processes, where it is sold, and which harms or decisions follow.

Regulatory gaps should be described precisely

A product may sit outside one regime and inside another. A low-risk wellness function may not be regulated as a medical device, yet deceptive claims, unsafe hardware, undisclosed data sharing, discrimination, or a data breach may still create liability. Conversely, a medically authorised implant can still raise unresolved questions about employment use, data ownership, software updates, company failure, access, and support after a trial.

A technology may feel liberating to one person and normalising to another. Restoring a capability, acquiring a new one, and being told that one’s existing way of being should disappear are morally different experiences—even when the hardware is identical.

Disability perspectives belong at the centre

Many neurotechnologies are developed for disabled people, yet priorities can be set by engineers, investors, clinicians, or families without equal power for intended users. For its States Parties, the UN Convention on the Rights of Persons with Disabilities requires respect for autonomy, non-discrimination, accessibility, participation, and inclusion.³⁰ Co-design should begin before the outcome is fixed: some people may prioritise communication accuracy, low fatigue, portability, repair, privacy, or keeping an existing mode of expression over a laboratory benchmark.

Respect also means allowing different conclusions. A person may welcome a restorative implant; another with the same diagnosis may refuse it or understand their disability as part of identity and community. Neither decision makes the person anti-science. The duty is to make both choices informed, practically possible, and free from abandonment.

Public opinion is specific to the scenario

General percentages such as “people support therapy but reject enhancement” conceal the wording, risk, population, and country. In a U.S. survey published by Pew Research Center in 2022, 77% favoured using brain implants to increase movement for people who are paralysed, and 64% favoured using them to treat age-related decline in mental abilities. Yet only 13% thought widespread use of a hypothetical implant that made information processing more accurate and faster would be good for society; 60% expected people would feel pressure to obtain it, and 57% expected the gap between higher- and lower-income Americans to grow.³¹ Those are findings about specified U.S. scenarios, not universal moral laws.

Transhumanism is a family of arguments, not a verdict

Some advocates see deliberate enhancement as a continuation of medicine, education, and tool use: a moral opportunity to reduce suffering and expand human capability. Critics worry about irreversible experimentation, commercial control, coercive competition, loss of diversity, and a society divided by access. “Playing God” and “progress is inevitable” are both shortcuts. Better debate compares concrete interventions, evidence, power structures, alternatives, and reversible steps.

Public trust is not secured by asking people to trust more. It is earned by making institutions worthy of trust—and making refusal survivable.

1. Define a valuable human purpose.Name the problem, affected people, desired everyday outcome, and non-technological alternatives. Do not begin with a device searching for a market.
2. Map benefits, burdens, and power.Identify who chooses, pays, profits, supplies data, performs unpaid support, can refuse, carries physical risk, and is affected if the system fails.
3. Build evidence proportionate to the claim.Use relevant comparators, meaningful outcomes, representative participants, long enough follow-up, independent replication, and public registration.
4. Design rights into the product.Use minimised data, accessible controls, safe defaults, explainable status, manual override, secure updates, portability, and exit without punitive loss.
5. Use independent, plural oversight.Include technical, clinical, ethical, legal, cybersecurity, disability, worker, child-rights, and community expertise—with conflict-of-interest disclosure.
6. Stage deployment.Set entry criteria, limits, stopping rules, monitoring, user support, complaint routes, and a sunset or renewal date before moving from study to wider use.
7. Measure life after the benchmark.Track function, learning, agency, mood, identity, sleep, relationships, inequality, rare harms, device dependence, maintenance, and outcomes after discontinuation.
8. Guarantee remedy and continuity.Provide correction, compensation where warranted, clinical care, safe explantation, data deletion where feasible, security support, and a plan if a company closes.

Red lines for institutions

• No hidden or compulsory neural measurement for ordinary work or education.
• No penalty for refusing non-essential neurotechnology or data collection.
• No punitive decision based solely on an uncertain mental-state inference.
• No claim that a narrow neural signal is unrestricted access to thoughts.
• No non-therapeutic optimisation of healthy children with invasive or irreversible technology.
• No sale of neural data or mental profiles behind vague bundled consent.
• No deployment without a secure update, maintenance, exit, and company-failure plan.
• No use of “ethics” as branding when independent audit and remedy are absent.

Before using or joining a study

• What exact ability or everyday outcome is expected to improve?
• Is this authorised treatment, off-label clinical use, research, wellness, or speculation?
• What is the best comparison—and how large and durable is the benefit?
• Which short- and long-term harms were measured, and what remains unknown?
• Could sleep, health, education, sensory care, rehabilitation, or another safer route address the problem?
• Can I stop, reverse settings, remove the device, keep ordinary care, and retrieve or delete data?
• Who sees raw data and inferences, and can they be sold, reused, subpoenaed, or combined?
• Who pays for maintenance, updates, complications, replacement, and removal?
• What happens if the maker changes terms, is acquired, ends support, or closes?
• Am I choosing this for my goals—or responding to fear, pressure, advertising, grades, or work demands?

A governance scorecard for organisations

Human intelligence can be precious: biologically grounded, developed across a lifetime, strengthened through learning, and capable of contributions that no purchased object or administrative system can simply replace. We should protect it from preventable damage, invest in its growth, and celebrate people who use it to enlarge knowledge and life.

But the value of intelligence is exactly why nobody should be coerced into changing their brain, reduced to a neural score, profiled by an uncertain inference, abandoned with unsupported hardware, or asked to trade mental privacy for education, employment, or care. Enhancement should remain in the service of the person—not the other way around.

The durable path joins two commitments. Raise the universal floor: health, education, rehabilitation, accessibility, safety, and freedom from exploitation. Keep the ceiling open: advanced learning, original work, therapeutic breakthroughs, assistive technology, and carefully governed ways of extending human capability. Give people room to think alone, people to return to, and institutions capable of supporting ideas once they emerge.

Progress is not merely a faster brain or a more powerful device. It is greater capacity joined to greater agency—more ability to understand and create, more freedom to choose, and more wisdom in deciding what power should never be used.