Our Heads in the Cloud

Transcript

Bridget Todd: Ok. So every day on my feed, there’s some new amazing thing AI has created. It’s going into all sorts of realms. And it’s pretty impressive. But here’s a question that gets to the very essence of AI and what it is. Can AI help us think? Can it actually be inside my brain? In a chip? Like all the hype we hear from neurotech startups rushing to develop the tech that lets people control their digital devices just by thinking about them. Without keyboards or touchscreens. Just how close are we to bridging the tech/brain barrier? The reality is: researchers are using tech to decode how we think. We just might be very close to unlocking the mysteries of the brain. This isn’t sci-fi anymore. Are we ready for this? I’m Bridget Todd. And this is IRL — the award-winning podcast brought to you by Mozilla Foundation with PRX. In this episode: Should we let AI into our brains?

Ian Burkhart: After losing the capabilities that I lost, I really was then looking for any way to restore that that is possible. And if that meant having a device implanted in my brain, that was that kind of risk-reward trade-off that I was comfortable with.

Bridget Todd: That’s Ian Burkhart in Columbus, Ohio. There are fewer than 75 people in the world who’ve had brain implants like his. Ian was among the first. Today, he speaks out for the rights of others to have the opportunity, too. It started with an accident that changed his life.

Ian Burkhart: So in 2010, I had just finished my freshman year of college, and went on a vacation trip with a few friends. We went to the Outer Banks in North Carolina to, you know, kick off summer vacation. And I was swimming in the ocean there. And I dove into a wave. And that wave then pushed me down into a sandbar, so essentially I was diving into the shallow end of the ocean. And I immediately, you know, hit and knew something was wrong because I couldn’t feel or move any part of my body.

Bridget Todd: The impact left Ian with a severe spinal cord injury. To this day, he can’t move his lower body or lower arms. While he was adjusting to his new reality and looking for solutions, Ian found out that he qualified for a clinical trial in 2014.

Ian Burkhart: I had, one of my physicians gave me a call and asked if I wanted to just come down to the research lab and look at the device that they were looking at using which is a muscle stimulation device on the forearm. It would restore hand and finger movements, but in order to do that it was going to be controlled by a brain-computer interface.

Bridget Todd: A brain computer interface — or BCI — translates signals from the brain into a computer.
There’s been a lot of buzz lately about Neuralink’s announcement that their first patient received an implant but research into BCIs goes all the way back to the 1960s when scientists implanted electrodes in monkeys to record their brain activity. In 2004 after decades of trial and error, a sensor was implanted for the first time into the skull of a human. The experimental device was developed by John Donoghue of Brown University. As a study participant for one year, Matthew Nagle - who had lost the use of his limbs - could operate a TV, check email, and open and close the fingers of a prosthetic hand. Ian Burkhart explains more about how the technology works.

Ian Burkhart: Instead of having your keyboard and mouse or touchscreen or trackpad, it’s coming directly from your brain. So in my case, that was an array that was implanted over the motor cortex. This device is about four millimeters square, and it has 96 little spikes that penetrate down into the brain, just one millimeter into the brain. And then through some fancy decoding and algorithmic software that is running, they’re able to infer what types of movements I was thinking about.

Bridget Todd: Ian’s BCI could only be used in the lab. So he would go there around 15 hours a week, between college classes, to train moving his hand and fingers.

Ian Burkhart: I could think about moving individual fingers, inflection or extension, and then combine those different fingers into multiple grips. That way I could reach out and pick up a bottle, take a drink. This system wasn’t, you know, really going to ever allow me to play the piano or type fully on a keyboard, but being able to allow me to grip onto objects and move them around is something that, you know, can really translate to an improvement in quality of life.

Bridget Todd: It worked. And the lab kept extending the clinical trial period. What was meant to be just 12 to 18 months, turned into seven and half years. And then, they took it out.

Ian Burkhart: At that point, I had the device for the longest that any human had ever had the device. And if there was any damage to the actual cortex in my brain, then that would lead to further paralysis in my arm. You know, it was definitely sad to know that, you know, now I don’t have the option of going into the lab and being able to connect with the system to have that increased hand function, but I also knew that it was time to be done and move on.

Bridget Todd: The number of people who had a BCI implanted or explanted, like Ian, is growing. In 2022, he launched a coalition for current and former research participants: BCI Pioneers.

Ian Burkhart: The biggest thing that we try to do is to highlight the experience of those who have been implanted with these devices in clinical trials, to really make sure that these devices are being created and developed in a way that can improve someone’s quality of life, not necessarily just pushing the science forward or being able to do something that’s cool, but really being able to do something that helps an individual.

Bridget Todd: Ian says the coalition wants BCIs to reach more people with disabilities. And with multi-million dollar implant startups recruiting volunteers with spinal cord injuries and ALS to test the technology, it seems possible. But Ian also wants that access to be on ethical terms. Like if a company goes out of business and the device is ‘bricked’ in someone’s head, who pays for the explantation —or if it can’t be explanted— how do you mitigate risk? In the coalition, they’ve talked about setting up a fund for these situations.

Ian Burkhart: We definitely feel like we can’t force companies to, you know, continue to develop a device that is no longer viable. But they still need to be able to support the individuals because at the end of the day, these are humans that they were able to use to improve their technology and they need to be able to then support those humans.

Bridget Todd: The other big area of concern is brain privacy or mental privacy- a concept sometimes described as neurorights. Ian says even neuro data about movements can be very sensitive.

Ian Burkhart: Maybe they’re not being able to read your thoughts. Some of these devices look like, you know, imagined handwriting and then each letter is essentially a keystroke on the computer or even imagined vocalized speech, so the motor movement that is required to make the speech happen. And for those devices in particular, we need to really make sure that there’s strong data privacy, or else it’s really just letting someone, you know, have a full record of everything you’ve ever wanted to say or write.

Bridget Todd: Do you have any sense of what a policy or a practice might look like that does allow you to protect your brain data?

Ian Burkhart: Yeah. You know, the most simple thought there is really just allowing agency and autonomy. And that the individual is the owner of all of that brain data and allow them to do what they want with it. We want to make sure that we’re not taking advantage of individuals like myself who, you know, are participating in these trials. And we don’t want to exploit those individuals who feel like they have nowhere to turn, besides having a chip implanted in their brain.

Bridget Todd: Any one of us could find ourselves in a position of needing to overcome challenges from a disability or illness. There’s also a lot of AI research happening with other kinds of neural tech. Like wellness headbands or earbuds that can detect neural activity. And using brain computer interfaces to give the ability to communicate back to people who have lost the ability to speak. So should we be worried about “mind reading”? Say, if I were thinking negative things about my government or boss? Stay with us, we’ll be right back. And we’re back. We’re surrounded by tech that collects volumes of personal data. Now attention is turning to neuro data. Neuro rights have really begun to capture the imagination of policy makers everywhere.

Maria Paz Canales: In the case of Chile, what happened a couple of years ago is that a group of parliamentarians and one particular parliamentarian got very interested in the challenges posed by neurotechnologies.

Bridget Todd: That’s Maria Paz Canales from Chile. She leads policy and research at Global Partners Digital, an advocacy group for human-rights-centered tech. In 2021, Chile became the first country to add neuro rights to their constitution, partly due to concerns over freedom of thought. At the time, Maria Paz was the leader of one of Latin America’s leading digital rights groups, Derechos Digitales. But she came out against the change.

Maria Paz Canales: So it sounds good, at first, hearing that you are creating a specific new set of protections for dealing with this very complicated issue, but this can create confusion in terms of the International Human Rights Framework to be flexible and adapt to the new technologies and the new challenges.

Bridget Todd: Some of the worst fears around neurotech are pretty dystopian, like governments using it for surveillance and social control. But freedom of thought is already a human rights issue.

Maria Paz Canales: A lot of dystopic examples that we can imagine again, in terms of the use of the technology, that can be specific of neurotechnology, but at the end, those are challenges that come with every wave of new and emerging technology that we are dealing with. The same with artificial intelligence, the same with autonomous weapons, and so on.

Bridget Todd: Maria Paz says that neuro data is sensitive but should be protected under general data protection frameworks.

Maria Paz Canales: It’s not that we are opposed to the rights. We are opposed to the fact that that specific set of rights need a separate definition. Sometimes there is a temptation of, like, going through to pick the new shiny thing, but not paying enough attention to other things that are more baseline but can provide a more sustainable environment.

Bridget Todd: By shiny new thing, Maria Paz means AI and neuro rights, which gets more headlines than regular old data privacy. Chile’s constitutional process sparked a chain reaction of policy discussions in other Latin American countries and in international organizations.

Maria Paz Canales: I mean, as being someone coming from the global majority, particularly from Latin America. I think that we are increasingly concerned about being regarded as a place in which companies can develop their business and extract the value of the technologies and extract also the data that is required for the improvement cycles of the technology and don’t share with the communities that have helped to create that value, any of the benefits that come with it.

Bridget Todd: As a longtime digital rights defender, Maria Paz sees how the hype around AI and neurotech also has an upside.

Maria Paz Canales: If people wonder what it is, question if it will be useful for them or not, who is controlling the technology, what are the rules, what technology is being used, all that is super relevant. And I think that as long as humanity in general, and each country and each society, we keep in that direction of questioning the evolution and integration of this technology in society, we will be in the safe place. So I want to believe that we can do it together.

Bridget Todd: Elon Musk has said that one of his hopes for brain computer interfaces is that memories could be saved and replayed in the future. Now, that’s a pretty cool idea, but what if there were a less invasive approach than a brain implant using AI to preserve memories? No surgery required.

Pau Aleikum Garcia: Synthetic Memories is a research initiative, actually, that recreates and preserves personal subjective memories. Generally, memories that are at risk of being lost. And we do that by transforming spoken and written depictions into images and videos.

Bridget Todd: Pau Aleikum Garcia is the founder of a research and design studio in Barcelona called Domestic Data Streamers. Their project, Synthetic Memories, is about using generative AI to visually reconstruct memories that were never documented. So, historic records of emotional truths.

Maria Carmen: [ORIGINAL] Bueno, m’emociono molt. M’emociono molt perquè cada vegada tinc més records. No sé què me passa, però és que cada vegada recordo més coses.

[TRANSLATION] Well, I get very emotional. I get very emotional because I keep remembering more and more things.

Pau Aleikum Garcia: Maria Carmen was one of the first participants of Synthetic Memories. We invited her at the studio. She’s a 94-year-old woman from Barcelona. And, we started with the question that we always do. What is your earliest memory? And I recall her saying, like straight to the point, saying, “Well, my earliest memory was when I was six year old and my mother used to bring me to another family’s place, another family’s house. And she will pay that family so we could enter their house, go through the stairs, up to the balcony.”

Bridget Todd: That balcony faced a prison where Carmen’s father was held as a political prisoner during the Franco dictatorship in Spain.

Pau Aleikum Garcia: And the only way they could see each other was from that balcony and the window of the prison. And that was her earliest memory of him, like between bars through that street. And when she explained that story, I asked her, like, “Would you like to have an image of this story?” And she said, “Yes, of course, like, this is important for me, this is relevant. This is something that I have explained to my family, but I have never been able to show it. And it is a very important foundational moment in my life.” And from there, we created, like, several images, 10 images. The first images, she was not, like, really connecting with it until we saw two specific images, and she was like, there was a very visceral reaction, she said, “Yes! That’s it, that’s my mother, I can see her haircut, her dress. It was like this. This is the balcony.”

Bridget Todd: Pau and his team used generative AI to create images based on Carmen’s memory. They also made a video, showing what it might have looked like from her father’s perspective, looking through the windows of the prison at Carmen and her mother. It’s sort of a poetic use of generative AI to spark true human connection.

Pau Aleikum Garcia: The goal was to depict her memory, and that was her memory — her subjective memory. It’s not a factual recreation of the past. It’s what she recalls from the past. Maybe the balcony was not like this. Maybe she was not on the right side of her mother. Maybe her mother had a different haircut. But this is how she recalls it. It’s actually part of why we decided to call the project Synthetic Memories, as these memories are very subjective. Because, and the more I read about memory and how it works, the more I realize about how bad we are at recalling things and understanding things

Bridget Todd: So like a story or a painting, the images capture the overall feeling of an event. For some participants, it can hold great meaning.

Pau Garcia: I think there is an intrinsic value on the fact that we can externalize memory, like explain something that goes beyond ourself. And more as we get old, that we start to see that at some point we will disappear and with us our memories, too. So, somehow, I have seen in Carmen — and, but in a lot of other people — a kind of a release, kind of a relaxation of seeing something that was only in themself, transform into something that can persist and can exist beyond their own lives. I remember this woman who told me, this feels a bit like finding your glasses in your head after searching for them everywhere. It’s like, it’s not a solution to the problem, but it’s a big relief, right? It’s a big relief to be able to see something that you thought you had lost.

Bridget Todd: At one time, Synthetic Memories opened a public office in Barcelona where anyone could book an appointment for a “memory reconstruction.” Sounds like a pretty good business idea, actually.

Pau Garcia: So this could be like, I think a terrible idea to transform it into a commercial project, yeah. I think it could drive it into a mercantilization of memories, which is already happening in so many platforms. But in this case, I think it could be even worse because we are tackling into memories that are very relevant, not only to yourself, but to your family, to your friends, things that somehow have defined who you are. So you can extract, I guess, a lot of metadata from the stories and understand very well people from that. So that’s a potential danger. That’s why I think it’s so important to anonymize all this information to, yeah, like, not leave it to a capitalization of how to transform these memories into money.

Bridget Todd: The synthetic memories are archived online, and have also entered the physical realm as part of a traveling exhibit to museums around the world.

Pau Aleikum Garcia: It’s a project that talks about who we were. It’s not a project about the future. And I think that’s one of the beautiful parts of this use of artificial intelligence. Most of the projects are now like really futuristic, looking at the future and trying to figure out what’s next. And we are trying to do something that is very old, that is just trying to recall better what happened in the past, trying to figure out new ways to understand ourselves. Like, where do we come from? I think that’s the value, a big part of the value of it.

Bridget Todd: So I imagine it won’t be long before I start seeing ads for brain tech on Instagram. Maybe not neural implants, but headphones and headsets that detect things like focus levels or dream states. Gadgets for wellness, productivity, maybe even fun. By helping us understand who we are, and how we think, AI should be used to help connect us to each other, not just to machines. That’s a no-brainer. Thanks for listening to IRL. For more about our guests, check out our show notes, or visit IRLpodcast dot org.