Are BCI's (Still) Overhyped?

Scroll through any tech feed today and you’d think we’re on the brink of telepathy. Neuralink is streaming chess moves from a paralyzed patient’s brain, Synchron is threading stentrodes into cortical veins, and every new inner-speech decoding paper gets spun as “mind reading.” At the same time, consumer EEG headsets and neuro-gadgets are flooding the market, promising focus, calm, and hands-free control. Behind the headlines, money is pouring in: nine-figure rounds for the usual suspects, new funds dedicated to “neural interfaces,” and total annual BCI funding crossing the billion-dollar mark. Between the venture capital, corporate partnerships, and glossy demo videos, it finally feels like BCIs are having their moment.

But as the volume of the hype rises, so does the pushback. In the academic world, review papers have started to push against the narrative, arguing that current BCIs are far less capable and far less scalable than the headlines suggest. Ethics and policy groups, from neurorights advocates to UNESCO and the OECD, warn about premature deployment, data misuse, and exaggerated claims that outrun both the science and the safeguards. And in mainstream media, pieces like the recent Guardian article take direct aim at “dumb transhumanist ideas” and billionaire fantasies, arguing that the stories being sold about neurotech may be doing the field more harm than good. Given this mix of acceleration and backlash, it’s worth asking a simple question: are BCIs still overhyped, and if so, where exactly is the gap between story and reality?

What We Mean When We Say “BCI”

At its core, a brain-computer interface is any system that senses activity in the nervous system and turns it into signals a machine can use, or sends signals back to change how the brain behaves. Two basic axes matter. First, invasive versus non-invasive: are you opening the skull or slipping something in through a blood vessel, or are you staying outside with electrodes on the scalp or skin? Second, monitoring versus modulating: are you reading out intent and state, or writing in pulses to nudge circuits? Under the same “BCI” label, you’ll find everything from soft EEG headbands marketed for meditation to high-density intracortical arrays sitting millimetres inside the motor cortex. Technically, they all qualify. Practically, they live in different universes of capability, risk, and regulation.

Precision Neuroscience's less-invasive ECoG grid (credit: Precision).

The road to today’s “neurotech moment” is much longer than most headlines suggest. In the 1970s and 80s, BCIs were lab curiosities: animal experiments, crude recordings, and early attempts to move cursors with EEG. Through the 1990s and 2000s, researchers built the first spelling interfaces and communication setups for people with ALS, alongside cortical implants that let a handful of volunteers control cursors and robotic arms. The 2000s and 2010s layered on DARPA programs, the Utah array, and the BrainGate trials: proof-of-concept systems that worked, but only in a few tightly supported patients. It’s only in the last decade that those threads have been pulled into venture-backed companies and pitched as platforms, not just experiments, giving the impression of an overnight wave that was, in reality, decades in the making.

Today’s landscape reflects that history. On the invasive side, you have intracortical arrays from groups like Neuralink, Paradromics, Blackrock and academic teams; ECoG grids laid on the cortical surface; and endovascular “stentrodes” from Synchron, delivered through blood vessels instead of open surgery. On the non-invasive side, there’s a crowded field of EEG and fNIRS devices aimed at research labs, rehab clinics, wellness consumers, and anyone who wants to control a cursor with eyebrow muscles. In terms of actual clinical scale, though, the numbers are humbling: we’re talking on the order of dozens of chronically implanted patients worldwide, not thousands, let alone millions. The brand footprint of BCIs has gone global; the installed base, for now, would still fit in a lecture theatre.

Not Everyone Is Buying the BCI Hype

If you listen to the headlines, BCIs are already halfway to telepathy. If you read the literature, the picture is more modest. Recent review papers criticize three recurring myths: that current systems can read arbitrary thoughts, that we’re on the cusp of general consumer deployment, and that brain interfaces will soon replace keyboards, mice, and touchscreens. The critiques continue when you get into the technical details of BCIs. Signals drift, electrodes move, tissue reacts. Most of the strongest results still come from a handful of labs working with a very small number of participants, under carefully controlled tasks. Models that look impressive in a paper can fall apart when you change the task, the environment, or the person wearing the implant.

That hasn’t stopped a parallel debate about mental privacy. Ethicists and neurorights advocates are right to worry about a future where neural data is treated like just another data feed, mined for hints about mood, intent, or political leanings. At the same time, most technical work today still lives in what you might call “weak mind-reading”: decoding specific, rehearsed tasks or constrained vocabularies, not peering into your private monologue on a random Tuesday. Policy, however, is deliberately trying to outrun the tech curve. Concepts like identity, agency, and mental integrity are being written into draft frameworks now, before anyone has a commercial system that could realistically test those boundaries.

Even if we take the most optimistic view of the science, turning BCIs into a business is its own hard problem. Implantable devices demand years of R&D, complex clinical trials, specialized manufacturing, and expensive follow-up care. Hospitals need trained surgeons, new workflows, and a clear answer to the question “who pays for this?” Reimbursement is murky, especially for small patient populations where the economics don’t look like a SaaS curve. Many market reports assume that if the tech works, adoption will follow. Yet in medical technology, adoption often slows down at the procurement and reimbursement stage, not because the underlying science is lacking.

Layered on top of all this is the story we tell about BCIs. Media coverage leans heavily toward enhancement, merging with AI, digital immortality, and sci-fi telepathy. The recent Guardian piece went straight at that culture, arguing that a set of transhumanist ideas-often pushed by billionaire backers-has raced ahead of what the devices can do. It’s not just that the promises are premature; they risk skewing priorities away from messy, patient-centred work toward cleaner, more cinematic narratives. In that sense, the “backlash” is less about hating on BCIs and more about asking a simple question: whose vision of the future are we actually building toward, and on whose terms?

There has been increasing criticism of Elon Musk (Neuralink) and like-minded individuals backing neurotech.

What BCIs Can Do Today

Strip away the slogans and you’re left with a smaller but more impressive set of realities. On the communication side, intracortical BCIs are now letting some people with paralysis type at close-to-conversational speeds in tightly controlled setups, outpacing many of the legacy eye-tracking and switch-based systems they used before. Motor BCIs can drive a cursor across a screen or move a robotic arm with enough precision to browse the web, send messages, or grab a drink, at least in lab and early home environments. Endovascular systems like stentrodes are showing they can record useful signals for years without open-brain surgery. For the small group of people who currently have access, these are not science projects, but the difference between relying on someone else for every digital interaction and being able to act on your own.

The reason serious people are optimistic about the next decade is that two curves are bending in the right direction. On the software side, AI-enhanced decoding is pushing beyond simple point-and-click, toward inner speech, more naturalistic movements, and models that adapt over time instead of breaking at the first sign of noise. On the hardware side, higher-density flexible arrays, fully implanted wireless systems, and minimally invasive delivery routes are chipping away at the risk and friction of getting signals out of the brain in the first place. Put together, these trends make a future with more capable, more stable BCIs feel plausible. But they are still early. Most of the headline-grabbing results live in single-digit patient cohorts, in carefully designed tasks, under the watchful eye of research teams.

That’s why it’s worth stating clearly what will not be announced in the next quarter. We are nowhere near general “mind reading” of arbitrary thoughts or memories. Routine cognitive enhancement implants for healthy people are still living in pitch decks, not in regulatory filings. And the idea that millions of people will ditch phones or keyboards for implanted hardware in the next five years ignores not just the science, but the reality of surgical capacity, cost, reimbursement, and informed consent. Those images mostly come from marketing, sci-fi, and the more enthusiastic corners of tech Twitter. The actual roadmaps, the ones that pass through ethics committees, regulators, and hospital procurement, point toward something slower, narrower, and, for the patients, more profound.

Where the Story Gets Ahead of the Science

If there’s one thing BCIs don’t lack, it’s ambition. The trouble starts when that ambition hardens into the aforementioned fantasies. Yet, part of the reason those visions stick is that the incentive structure practically demands them. Founders building hard tech over ten-year timelines have to sell a future big enough to justify the risk, the burn rate, and the fact that, for a long time, there is nothing to show but demos and papers. Venture investors, in turn, want large total addressable markets and paradigm shifts, not a few dozen patients per year.

Media outlets know that “implant helps locked-in patient send emails slightly faster” will get a fraction of the clicks of “brain chip lets man play video games with his mind.” Some of this is simply long-term aspiration expressed clumsily. Some of it is marketing that crosses into misrepresentation. Either way, when the gap between promise and delivery becomes too wide, it doesn’t just dent one company’s reputation. It risks souring regulators, clinicians, and the public on the whole category of innovation.

That’s a shame, because in the places where BCIs genuinely shine, they’re still underrated. For someone who is locked in or severely paralysed, gaining a reliable, high-bandwidth channel to communicate or control a device is not a lifestyle upgrade but a serious lifeline. In certain rehab or neuromodulation setups, BCIs create feedback loops that simply weren’t possible with older tools, helping patients relearn movements or regulate symptoms in more targeted ways. These are narrow use cases, but their impact per person is enormous. And the work done to support them-better decoding algorithms, more robust hardware, smarter clinical workflows-spills over into the rest of neurotech. If there’s a place where optimism belongs, it’s here.

Less Magic, More Medicine

So, are BCIs still overhyped? In the public imagination and in a decent chunk of media coverage, yes. The gap between what most people think these systems can do and what they actually deliver is wide. But that doesn’t make the field a scam, bubble, or dead end. It just means that some expectations are misaligned with the physics, biology, and timelines of the tech.

Right now, BCIs are over-marketed and undervalued at the same time: oversold as a path to mind uploading and human-AI fusion, underappreciated in their very real strengths for small groups of patients. But, if we stop expecting BCIs to upload our minds and start judging them by how well they help people live, communicate, and move, we might find that the real story is less magical and far more important.