Artificial intelligence has come to Being able to improve our day to dayand this is what a team from the University of California wanted Improve the quality of life of people with disabilities.
Machine control with the mind. Something that a priori can be taken from a science fiction film, but that neurotechnology laboratories want to lead to reality. The problem is that precision and accessibility remain a real challenge for all engineers.
In this way, the Cerebro-Order interfaces (BCI) right now they require very complex surgery for Be able to install electrodes in the brain To achieve great precision. But now this will change, According to a study Published in the prestigious magazine Nature Machine Intelligence which presents an advance that could change the rules of the game: an AI that acts as ‘co -pilot’ of a non -invasive BCI that improves its performance.
A team that has already been tested. The system, developed by Jonathan Kao’s team at the University of California, has allowed a man with partial paralysis to control a robotic arm to perform complex tasks, something he could not do with conventional BCI technology.
A co -pilot of AI to read the mind. Non -invasive BCIs, which are placed on the scalp to capture the electrical signals of the brain, are safer and more accessible, but also less precise. The “noise” and the weakness of the signals make it difficult for the system to decipher the user’s intentions exactly.
This is where artificial intelligence enters. Instead of trying to decode each nuance of brain activity, Kao’s system Use an AI that collaborates with the user. And this is how Kao himself explains to Nature:
These co -pilots are essentially collaborating with the user of the BCI and trying to infer the objectives you want to achieve, and then help complete those actions.
In this way, artificial intelligence learns to integrate the intention of the user himself, needing less direct brain information to execute a precise command. This creates a shared autonomy system, between the human and the machine itself.
Results that speak for themselves. To test this system, the researchers conducted two key experiments. First, they asked four participants (one of them with paralysis) to move a cursor on a screen to a goal. With the traditional BCI they were achieved most of the time.
However, by activating the co -pilot of AI, they all completed the fastest task and with a much greater success rate, improving the performance up to four times.
The fire test. This was done with a real robotic arm. The task was to take colored blocks and move them to specific points on a table. The participant with paralysis could not complete the task using the non -invasive BCI alone. However, after activating the AI, its success rate shot up to an amazing 93%. Participants without paralysis were also remarkably faster with the help of AI.
A view to the future. This approach opens a promising door to improve the quality of life of people with motor disabilities, offering a powerful tool without going through the operating room. Zhengwu Liu, an engineer from Hong Kong University who did not participate in the study, qualifies the idea as “a good way to achieve a more powerful man-machine hybrid system.”
There is an ethical challenge. The advance is not free of controversies, and as the neurologist Mark Cook of the University of Melbourne points, it is necessary that the shared autonomy does not occur “at the expense of the user’s autonomy”, since “there is a risk that the interventions of the AI annul or misunderstand the intention of the user.”
Kao himself recognizes this challenge. In fact, he reveals that the participants in the study did not like the versions of the system in which the AI had too much control, such as when he could decide the trajectory of the robotic arm itself. Finding that right point where the AI assists without dominating will be key to the future of this technology, which promises to continue evolving with more sophisticated co -drivers and, even, its possible integration into the BCIs surgically implanted.
Images | Ubidium Beach
In Xataka | Hong Kong faces a huge public health crisis by the Chikungunya virus. Its solution: an army of robot dogs
GIPHY App Key not set. Please check settings