In John Scalzi’s sci-fi novel Old Man’s War, soldiers receive the BrainPal, a neural interface that boosts thinking and connects them to computer assistants. Sounds far-fetched? Perhaps not for long.
Military neuroenhancement may be closer than we think. An AI-assisted closed-loop brain-computer interface, or AI-BCI, could be a breakthrough tool to upgrade warfighters. But while the technology may offer real benefits, it also raises major neuroethical concerns: it may weaken users’ autonomy and shift human reasoning toward algorithmic logic.
AI-BCIs: From Therapy to Enhancement
Implantable AI-BCIs are used therapeutically for epilepsy, Parkinson’s disease, and treatment-resistant depression. Yet in healthy users, similar non-invasive systems could enhance attention or reduce fatigue and stress.
Think of an AI-BCI as a smart fitness tracker and pacemaker in one: it monitors and stimulates neural activity, adjusting mental states. The device records brain signals, uses AI to detect, for example, cognitive overload or anxiety, and stimulates the brain to correct its activity. It resembles “a tag-team of spy and sleeper agent.” Once the “spying” unit detects an anomaly, it “wakes up” the stimulation module.
One psychiatrist compared an AI-BCI to an electric bicycle: “When someone’s pedaling but having difficulty, the bike senses it and augments it. We’ve made the equivalent of that for human mental function.” While compelling, the analogy is imperfect – unlike an e-bike, an interface operates without the user’s full awareness of when and how it modulates their brain activity. And unlike caffeine or stimulant drugs, whose effects are usually consciously felt, it works below that level.
The Military’s Interest in Neurotech
U.S. military researchers have long explored brain technologies, and the Defense Advanced Research Projects Agency has supported projects showing that AI-BCIs may help regulate mental health and enhance memory. Noninvasive solutions are also emerging rapidly, including ultrasound and other forms of brain monitoring and stimulation.
Commercial start-ups are advancing wearable neurodevices, such as headbands and headphones, for wellness, entertainment, and productivity. For example, the Somnee headband tracks and modulates brainwaves to improve sleep quality.
The military is also moving toward monitoring soldiers’ brains – a prerequisite for neuroenhancement. For instance, the U.S. Air Force partnered with Neurable to track cognitive fitness and attention with EEG headphones, while Aptima is developing an AI-enabled helmet system to monitor helicopter pilots’ cognitive states. The next stage will integrate these neural scanners with brain-zapping capabilities.
The benefits seem clear: boosted cognition and improved emotional control could help soldiers handle fatigue, operate efficiently, keep pace with accelerating human-machine interactions, and prevent errors. But neuroenhancement is not morally neutral. Its ethical challenges should therefore be addressed before such solutions become operational.
Autonomy Under the Algorithm
The first issue is human autonomy, and three approaches are useful here: traditional, experiential, and relational.
In the traditional view of autonomy, individuals are autonomous when they act intentionally, understand what they are doing, and remain free from undue control. AI-BCIs may not directly threaten intentionality, since they cannot implant motives or goals. But they could challenge users’ understanding and meaningful control over their choices and conduct. Because closed-loop interfaces operate below consciousness, soldiers may not fully understand why they decide or act as they do. Consider autocorrect. It fixes typos but sometimes changes the intended meaning before we notice. AI-BCIs could create a cognitive autocorrect mode. Warfighters may not know what the algorithm inferred from their brain data or how it adjusted their mind. Even with informed consent, a gap remains: you can agree to use the device, but can you truly understand its moment-to-moment influence on your psyche? And there is the problem of non-control. AI-BCIs do not just support cognition; they can filter, prioritize, and narrow options. Soldiers may still feel in control, even though the BCI has quietly structured their mental activity.
The experiential framework of autonomy emphasizes the lived sense of agency: the feeling that I am the one who initiates, controls, and owns my actions. But when decisions are co-managed by an adaptive system, users may feel alienated from their behavior. Was that truly my judgment, my decision – or was I pushed by the system? Some patients undergoing deep-brain stimulation have reported feeling “like a machine” or an “electric doll,” detached from themselves, as if their actions were no longer truly their own. BCI-enhanced soldiers could experience similar estrangement.
The relational approach views autonomy as embedded in interactions with others and our social environment. Yet in the age of AI, “others” are not only humans; they are also smart machines. If a soldier’s cognition is partly co-produced by an AI-BCI, agency and responsibility become harder to locate. If the system misreads neural signals and contributes to wrongful action, who is accountable? The soldier? The commander? The manufacturer? The AI? As cognition is shared between biological and digital intelligence, accountability may diffuse.
Ultimately, autonomy may need to be redefined as a hybrid human-machine concept.
Human Reasoning
An even deeper concern is that AI-BCIs may alter the nature of human reasoning. Our thinking cannot be reduced to statistical optimization. We do not reason solely through pattern recognition or prediction. Human judgment is contextual, embodied, emotional, and often abductive: we move from fragments of reality to hypotheses and interpretations. AI operates differently. It is inductive and predictive. It identifies patterns in past data to optimize future outputs. AI models prioritize what is statistically likely, not necessarily what is contextually wise or morally right.
This distinction matters in combat. War involves uncertainty, chance, and friction, so military decision-making depends on more than processing speed and statistical calculations. Soldiers need flexibility, intuition, and the ability to respond to what does not fit prior patterns. AI-BCIs, however, may favor what has already been learned as optimal, pushing cognition toward logical consistency.
Recent research illuminates this risk. Steven Shaw and Gideon Nave build upon the dual-system theory of cognition, in which System 1 represents fast, intuitive thinking, while System 2 involves slow, deliberative reasoning. They introduce System 3 – external artificial cognition. When humans interact with AI, they often experience “cognitive surrender.” Automation bias, our tendency to over-rely on technology, often leads people to adopt AI outputs without critical evaluation. In their study, participants followed incorrect AI advice nearly 80 percent of the time. Thus, System 3 may override both intuition and deliberation. And when integrated with the brain through AI-BCI, it may no longer remain merely external. System 3 could become part of cognition itself, subtly steering human reasoning while feeling like one’s own judgment.
Who Is Really in Command?
AI-enabled interfaces may offer tangible military benefits. But neuroethics requires us to ask a deeper question: What kinds of subjects do they produce?
Without robust testing, clear safeguards, transparency, accountability, and meaningful human oversight, military neuroenhancement may erode the very human element it claims to strengthen. If algorithms begin managing soldiers’ cognitive functions without their full awareness, military AI-neurotech could undermine the very human agency.
Lukasz Kamienski, PhD, (@LukaszKamienski) is a Professor of Security Studies at Jagiellonian University in Krakow, Poland, and a Fulbright Visiting Scholar at the Center for Ethics and the Rule of Law at the University of Pennsylvania.