Chip in the head: When computers learn to understand thoughts

Chip in the head When computers learn to understand thoughts As of: October 31, 2026 • 10:17 a.m

What sounded like science fiction for a long time is now reality: chips in the brain help paralyzed people control computers with their thoughts. Researchers hope for new therapies, ethicists warn of risks.

26-year-old Michael Mehringer sits in his wheelchair in a laboratory at the Technical University of Munich Hospital. A USB port peeks out of the back of his head. This adapter is intended to give Michael something back that he didn't have for a long time: independence.

"I've been more or less trapped in my own body for ten years now," says the young man, describing his situation. At the age of 16, Michael had such a fatal accident in a motorcycle accident that since then he has been completely paralyzed from the neck down and has had to rely on his parents' help.

When he became aware of a study by the Technical University of Munich last year, hope began to arise in him: hope for a bit of freedom. For the study, electrode chips are implanted into his brain in a complex operation. This marks the beginning of a long, arduous journey for Michael, but today he says: "I'm glad I took this step."

The adapter on the back of Michael's head is part of a brain-computer interface, the so-called BCI. The brain-computer interface processes signals from 256 tiny electrodes from Michael's motor cortex, the area of ​​the brain that controls movements.

“They are so small that each individual microsensor can record the electrical flashes from a single nerve cell,” explains Simon Jacob, head of the AI ​​and robotics in neurosurgery working group at the Technical University of Munich in the ARD knowledge documentary “Chip in the head. How brain implants are changing medicine.”

These signals are then transmitted to a computer via the USB port. Using an AI specially developed for this purpose, the researchers are trying to decode the nerve impulses. "The artificial intelligence learns this pattern. This means that when it sees an electrical signal in the brain, it can predict what Michael was thinking," says Jacob, describing the system.

New AI determines the time of stroke twice as precisely as previous methods

The computer will gradually learn to interpret the impulses from Michael's motor cortex and convert them into movements. The goal: In the future, Michael should be able to control a robot arm using just his thoughts.

Jacob is convinced that people like Michael would make a decisive contribution to basic research with their participation: "This is the technology of the future for many, many diseases for which we previously had no offer."

Brain-computer interfaces are still at the beginning of their development. But what sounds like science fiction has long been reality and technology is developing rapidly. The Freiburg company CorTec has developed an implant that can not only read signals, but also send them to the brain and thus stimulate nerve cells.

“We get the brain to be more capable of learning when it comes to strokes, so that physiotherapy works better,” explains co-founder Martin Schüttler. The first tests are promising. In the USA, a stroke patient who was implanted with a CorTec-BCI in a study in 2025 was able to move his paralyzed arm again.

Elon Musk's brain chip company Neuralink pursues one goal: the symbiosis of humans and AI. more

These groundbreaking developments give hope. But the possibilities of neurotechnology are no longer only recognized in a scientific context. Elon Musk's company Neuralink plans to put brain chips into mass production in the long term and, according to Bloomberg, will carry out 20,000 implantations annually from 2031. Musk's vision: a fusion of humans and AI.

However, Neuralink does not yet have approval from the US health authorities. A BCI developed in China, on the other hand, has already been officially approved for commercial use. The implant is intended to help paralyzed people move a hand using an exoskeleton - a type of wearable support.

Neurologist Jacob attributes the current development boost primarily to advances in artificial intelligence: "The really big shift that we saw was actually in software, i.e. in artificial intelligence." AI is getting better and better at reading out meaningful patterns from large amounts of data.

As early as 2023 - studies from Osaka University and Berkley University show - researchers were able to partially reconstruct images and music from brain data using AI. Researchers at Stanford University recently even managed to partially decode the inner monologue of paralyzed patients in an attempt. To prevent unintentional mind reading, the system only responded to a mental password.

This attempt at mind reading can help people who can no longer communicate. But it also raises questions: How free are our thoughts really? "At the moment when I can, on the one hand, read and, on the other hand, manipulate things, there is also a huge field for abuse in things like mental privacy," says Judith Simon, Professor of Ethics in Information Technologies at the University of Hamburg.

The neuroethicist Marcello Ienca is a professor at the Institute for Ethics in Medicine and Technology at the Technical University of Munich. He currently sees the greater risk not in implanted chips, but in neurotechnologies for everyday life, such as wearables such as headbands or caps that can measure brain signals. These could become relevant in education or the world of work, for example: "If neural data are used, for example, to measure attention, fatigue or stress, the question immediately arises as to whether employees can actually freely decide to reject such systems," says Ienca.

There are currently no brain implants that can demonstrably increase general intelligence in healthy people. But if technologies at some point enable cognitive advantages, according to Ienca, a question of justice arises: "Who has access to them? A society in which cognitive abilities can increasingly be improved commercially could increase existing social inequalities."

Think of a word and the computer says it. more

So far, these scenarios are still partly theoretical. Nevertheless, we must already ask ourselves today: “How we protect the measured data, i.e. who has access to it and what is done with it,” says neurologist Jacob. "So we shouldn't block these technologies, but rather develop them in an ethically responsible manner."

Development happens one way or another, so “we have to get used to the things that happen,” says Simon, who provides scientific advice to the EU Commission on questions relating to AI. "But that doesn't relieve us of the responsibility to think about how we can regulate such things wisely." Our mental privacy needs to be protected: “It is not without reason that it is said that thoughts are free.”

Chile already has brain data protection rights enshrined in its constitution. There is still debate in Europe about how mental privacy can be protected in the future. The UN warns of the risks in a current report. Neuroethicist Ienca also points out that technological developments can make gaps in protection visible: "Our thoughts and mental processes should not become another resource that states or companies can access without very strict requirements."

In the case of Michael Mehringer, the debate is different. The researchers at the Technical University of Munich need to know what is going on in his head in order to help him. And indeed: after months of training, Michael achieves amazing things. He moves a cursor on the screen using just his thoughts and shortly afterwards even a virtual hand. “When something like this works, the enthusiasm is naturally high,” says Michael happily. For people like him, the new technology already means one thing above all: hope for more freedom. “My dream would be to regain 100 percent independence and relieve the burden on my parents.”

This topic in the program: The First | Chip in the head - How brain implants are changing medicine | Oct 26, 2026 | 11:05 p.m