Should We Be Allowed to Sell Our Brain Data?

As physical AI takes off, it’s easy to find examples of workers in low-resource countries wearing head-mounted cameras and performing everyday chores to help robots gather labeled data from the physical world. This echoes what fueled the computer-vision boom a decade ago, when ImageNet (Fei-Fei Li) supplied the labeled data that modern vision models were built on. So why wouldn’t the same thing happen to neurotechnology once these devices are commercialized?

One question has been on my mind since a recent conversation with a friend, Caleb McKinney: why don’t companies like Neuralink openly share anonymized brain data collected from patients?

Brain–computer interfaces (BCI) have the potential to become one of the most transformative technologies humanity has ever built. Yet we seem to be repeating the pattern we’ve already seen with social media and AI: data becomes proprietary, companies compete to accumulate it, and progress is gated by ownership rather than collaboration.

This raises difficult questions. If a patient consents to a BCI study, who truly owns the resulting neural data? Is it the patient, because those signals originate in their brain and represent their experiences? Is it the company, because the data was collected using its device and infrastructure? Or should anonymized neural datasets be treated more like a public scientific resource that benefits everyone?

Unlike many other forms of data, neural recordings are extraordinary. They contain compressed representations of perception, intention, movement, and cognition—and perhaps, one day, much more. If properly anonymized and shared responsibly, these datasets could dramatically accelerate neuroscience, AI, rehabilitation, and accessibility research. Instead, there is a strong incentive to hoard them as competitive assets, much as behavioral data became one of the most valuable resources for social media platforms.

How about privacy? Even “anonymized” neural data may not stay anonymous. Because these signals encode behavior, they could potentially be aligned or matched with the vast behavioral consumer datasets that companies already collect—keyboard typing patterns, speech and speaking rhythms, gait, and other biometric signatures. If such cross-linking becomes possible, neural recordings could be re-identified and tied back to specific individuals, turning a supposedly de-identified dataset into one of the most personal fingerprints imaginable.

This makes me wonder whether we are importing the wrong incentive structure into one of the most sensitive technologies humanity has ever developed. Some level of competition may be necessary, but should proprietary data ownership be the primary incentive in this field? Or should BCI evolve under a fundamentally different model of collaboration? That worries me.

One reason I’m excited about BCI is that the field is still young enough for us to shape its culture before it matures. I hope we don’t simply copy the incentives that emerged in today’s AI and social media ecosystems. BCI is not just another software platform—it interfaces directly with the human brain, and that demands a far higher standard of ethics, transparency, trust, and long-term responsibility.

Perhaps progress in BCI shouldn’t be driven primarily by who owns the largest proprietary dataset. Perhaps it should be driven by new models of collaboration among academia, industry, patients, and society—where privacy is rigorously protected but scientific progress is shared whenever possible.

I don’t claim to know the answer, and I’m genuinely curious how others think about this:

• Who should own neural data?
• How much should be open?
• Can we build a BCI ecosystem that prioritizes collaboration without sacrificing innovation?
• Should the incentive structure itself be reimagined for technologies that interface directly with the human brain?

I’d love to hear perspectives from researchers, clinicians, entrepreneurs, ethicists, and especially patients.