Marco Giacomello.Mens et opusITWork with me
Kairos·Essay·Neurolaw·Habeas Mentem·Legal design

The people with the implant write the rules

Six people living with an implanted brain-computer interface have written their own rules for research. They ask for something the law takes for granted: consent that is understood, not just signed.

17 November 20263 min read7 sources

On 10 November a document came out that won't make much noise. It's called OUR BCI. It was written by six members of the BCI Pioneers Coalition, a community of people living with an implanted brain-computer interface, together with two researchers from York University and the University of Ottawa. It is version 1.0 of ethical recommendations for implanted BCI research. Thirteen chapters.

I have read many guidelines on neurotechnology. Almost all of them were written by committees, agencies, lawyers. Not this one. It's the first time I've read rules about the mind written by people whose minds are connected to a device.

What they ask for

They ask for concrete things. That brain data be treated as the participant's personal information, with researchers acting as stewards. That access to your own data always stays open. That no confidentiality agreement stops you from talking to your doctor, your lawyer, a regulator. That before surgery it is clear who owns the hardware, the software, the decoder built on your brain. And that a company's property interest is not enough, on its own, to decide whether an implanted component is removed or switched off.

This is not a conference hypothetical. Ian Burkhart, one of the authors, had regained movement in his arm thanks to an implant. Then the device was removed, and he lost that movement again. In Australia, Rita Leggett, implanted to treat epilepsy, had her device taken out against her will after the company that made it went under. The contract ends. The body stays.

Consent as a conversation

The part that struck me most is about consent. The document doesn't talk about forms. It talks about a process. It asks for materials without jargon, accessible formats, time for questions. It asks researchers to check, over several meetings if needed, that participants understand and remember what they signed. It asks for regular check-ins by someone outside the research team. And it asks for consent to be reconfirmed when something material changes: new uses of data, the length of agreements.

Among the testimonies there is one line worth the whole document: "Consent should be a conversation that continues as the study changes."

It is the opposite of what I call, in Habeas Mentem, the liturgy of consent: a ritual that produces a signature, not understanding. The law presumes that whoever signs has understood. Here it is the signatories who say that presumption doesn't hold. And they ask for understanding to be verified.

I can say it with our own data too. In the chapter I wrote for Mente e Tecnologia we measured how hard legal texts are with the LX Complexity Score, in an exploratory study of 100 people using EEG and fNIRS. None of the documents on brain-computer interfaces fell below the experimental threshold of 45. Syntactic density 72 out of 100, semantic distance 68. In plain words: the texts asking for the most delicate consent are among the hardest to understand.

From recommendation to rule

OUR BCI is not binding. The authors know it and say so: it is a first step towards a code that ethics boards, funders and regulators can enforce. Some points remain vague. Maintenance guaranteed "for a reasonable time period" after a trial is not yet a rule.

But the method is right. Whoever designs the study should prove the participant has understood, not presume it. That is the reversal of the burden: responsibility shifts to whoever builds the choice environment.

In Europe, the GDPR treats health data as a special category. In the US, several states already have neural data rules, while at federal level the MIND Act has been introduced but not passed. None of these texts, though, says clearly what happens to the person when the study ends, the company closes, the software stops being updated. The document asks for technologies that can withstand commercial discontinuation. That should be a design requirement, not a wish.

Clarity, not simplicity. Nobody is asking to dumb down a neurosurgery protocol. The ask is to design it for the person who has to live it. With Neuro Legal Cortex this is exactly what we work on: measuring whether a text is understood, instead of taking it for granted. Understanding can be measured. The mind is not to be read.

Personal reflection, not legal advice.

Newsletter

Two ideas a week, straight to your inbox

  • Notes on law, mind and technology
  • Open challenges, before they hit the site
  • Early access to new products

← All ideas