Tetiana Aleksandrova is Founder & CEO of brain-computer interface company Subsense Inc.gettySurgical brain-computer interfaces (BCIs) were an important step forward for conditions with few treatment options. Even if this technology was only for a select few, it proved that neural interfaces work. Those early systems also shaped what the field came to accept as normal. We are now in a position to question those assumptions—and we should.In my previous article, I argued that what holds BCI back is not biology, but rather what the industry as a whole has decided to focus on. The same is true for safety and ethics.The Ethics Conversation Starts Too LateMost BCI ethics conversations are centered on privacy, consent, autonomy and regulation. Those concerns are real, and I take them seriously. But by the time they reach the table, the most consequential decision has already been made: The field has chosen the system it will build. That choice carries its own ethics.If the premise is that a BCI requires a permanent surgical implant, there are already ethical problems baked into the architecture. These problems are not things you can manage later by bolting on a governance layer. Yet, the field often talks as though risk and performance are inseparable when it comes to BCI technology—as though a better signal must require a more invasive system, and a lower risk must mean weaker performance. Once that assumption is accepted, the ethics debate shrinks to a single question: Is the benefit large enough to justify the harm to a small group of patients?In a clinical trial, that may be the right question. But I believe it is the wrong question for a field that keeps describing itself as building something for broad human use. The deeper question we need to ask is this: Why is the trade-off between invasiveness and performance treated as a law of nature in the first place?None of this is a condemnation of the pioneers in this field. For some patients today, an invasive interface is genuinely the right call, and the people who built those systems created both hope and the proof that BCI works. Respecting that is not the same as accepting that their constraints must be everyone’s constraints forever.Safety Is What Happens After The ProcedureSafety in BCI is not only about the procedure. It is about what happens after. A device can be placed flawlessly and still run into long-term trouble where tissue meets hardware. Performance depends on biology: tissue response, stability and what happens over months and years.This is why much of the safety discussion still feels too forgiving to me. Neurotechnology has produced genuinely remarkable demonstrations—thought-controlled cursors, speech decoding and robotic control. Those achievements matter. But a breakthrough under controlled conditions does not settle questions of long-term tolerability, scalability or access beyond a select patient group. A system can be extraordinary once and still rely on an approach that is too hard on the body over time. Technology does not become ethically convincing because it works in a demo. It becomes convincing when there is a credible path to making it safe, durable and practical for many.Reversibility Is An Ethical Property, Not Just A Technical OneThis is where safety and design meet directly. The questions everyone cares about—consent, autonomy and control over one’s own neural data—are not abstract principles you can satisfy with a policy. They are shaped by what the hardware physically allows.A permanent implant asks a person to make a one-way decision. A temporary, reversible interface does not. Reversibility means consent can stay genuinely revocable, the cost of being wrong is lower and a person keeps control over whether the interface remains part of their life. An architecture that can be undone treats autonomy as something built in, not merely promised.When pursuing reversibility at Subsense, we made sure to pressure-test our foundational assumptions before committing to a direction, and I would urge my fellow leaders to do the same. Before choosing your path, consult people from across the industry, including leaders in chemistry, physics and nanotech, to surface what the field had accepted without examining it. Treat a lower burden on the body as part of your technical target, not a compromise to be made up for elsewhere. Layer these safety goals into your architecture decisions from the start, not after performance has been proven.A Higher Bar For The Whole FieldI am not interested in claiming that one company cares about ethics and others do not. I am interested in changing what the field believes and competes on as a whole.Neurotechnology has embraced the mindset of a conservative medical device world that is comfortable managing old limits. But BCI is precisely the area where staying inside that mindset may be the costliest mistake you can make. Once a field decides that a certain level of harm, complexity and exclusivity is simply the price of entry, it shrinks its future.In my view, technologies tend to expand only after they lower user burden. Continuous glucose monitors changed diabetes care by reducing friction. Wearables expanded monitoring by reducing invasiveness. Neurotechnology will likely head in the same transition—and the companies that want to lead it must champion safety, durability and low impact on the body.The real test of safety and ethics in BCI is not whether a company can justify risk in its most dramatic case. It is whether the field is willing to keep searching for architectures that make that risk less necessary. A BCI that works once but becomes harder to tolerate over time is not a mature system. It is an unfinished one.If we build around safety and reversibility from the beginning, BCI has a chance to become something for millions, rather than something for just a few. That is what safety by design means—not a constraint we accept and then learn to regulate, but the first thing we design around.​​Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?