Research on the development of artificial intelligence brings surprising results. Texas neurosurgeons are conducting experiments to facilitate what we colloquially call mind-reading. The technologies developed there would in practice make it easier for non-speaking people to communicate. CNN TV footage.
Neuroscientists at the University of Austin in Texas have made a breakthrough in research. They found a way to translate brain activity into written language with the help of artificial intelligence. This month, researchers published a paper in which they describe how a group of volunteers listen to audio recordings while an artificial intelligence device scans their brainwaves. The effect of these activities is a verbal record of what the subjects thought about. The same technology is used by Chat GPT. Thanks to this, the machine “read” what the volunteers were thinking only on the basis of the analysis of their brains.
– It’s amazing. Based on the observation of how blood flows through the brain, artificial intelligence was able to convert someone’s thoughts into words, explains Donie O’Sullivan from CNN.
To test the effects, Professor Alexander Huth subjected a CNN reporter to a CT scan while listening to an audiobook of The Wizard of Oz.
First, the professor scanned an image of the brain. Then he checked what blood flow through the brain looked like while listening to an audiobook. He later showed how my brain interpreted the recording. – We see a record every two seconds. The person listens to the recording. We run the data through our decoder and then process what the listener receives, explains Jerry Tang, a graduate student at the University of Austin.
Are we going to live in a world where all you have to do is walk past someone on the street and they can put something to your head and know your thoughts? – We’re a long way from that. It may never work, but we can’t completely rule it out either. With the current state of knowledge, this will not be possible in the next few decades. The real purpose of this technology is to help people who cannot speak. Then they don’t have to undergo neurosurgery, Professor Huth explains.
Artificial intelligence is used to create deepfakes, but it can also recognize them. “We have to be very careful”Jacek Tacik/Fakty TVN
A model used to decipher human thoughts
Jerry Tang explained how the Chat GPT model is used to decipher people’s thoughts. This type of artificial intelligence is based on the ability to analyze the content of millions of web pages. On this basis, the system remembers the construction of sentences. Thanks to this, he checks how people speak and think. “Thanks to the GPT, we have significantly improved our forecasts,” he points out.
This technology is not only useful for listening to audio recordings. Professor Huth also showed what was recorded when watching a film without sound. The technology quite accurately described the watched fragment.
“We believe that the effects of brain decryption need to be carefully monitored. We should also start working on protecting the privacy of our thoughts. The point is to determine how and for what the data obtained in this way can be used, emphasizes Jerry Tang.
Main photo source: CNN