For years scientists have been trying to make mind reading a reality. because the applications are boundless.
Already, wires between a prosthetic limb and the brain allowed someone to control an artificial hand with their thoughts and even feel what it held.
Now, computers can read a person's thoughts without any wires. Researchers did brain scans using fMRI, functional magnetic resonance imaging and artificial intelligence was able to interpret and decode a person's thoughts.
The AI is tapping into how ideas, semantics and meaning are generated. In this study, scientists had volunteers spend sixteen hours in an fMRI machine listening to podcasts.
Since fMRI can "take" images milliseconds apart, they can show how the brain responds to this verbal input. Images of these responses were fed into a neural network language model driven by an early version of the ChatGPT AI technology.
The AI learned from the fed responses how to predict what the brain would think about the spoken words. Volunteers were then placed back in the fMRI machine and listened to a new podcast.
It was able to decode the general gist of the story and in another test, it could decipher the thoughts of people who watched a silent movie.
This shows that the fMRI and AI were reading thoughts and not just recognizing speech. This will change the life of someone unable to communicate.
It's remarkable work even if we can see the downsides such as rogue governments reading its citizens' thoughts. Technology is often can be double-edged sword.
You can now hear additional episodes on many of your favorite podcast providers - visit Buzzsprout to subscribe.
More Information
Scientists Use Brain Scans and AI to 'Decode' Thoughts
Scientists said they have found a way to use brain scans and artificial intelligence modeling to transcribe "the gist" of what people are thinking, in what was described as a step toward mind reading...
Semantic reconstruction of continuous language from non-invasive brain recordings
A brain-computer interface that decodes continuous language from non-invasive recordings would have many scientific and practical applications. Currently, however, non-invasive language decoders can only identify stimuli from among a small set of words or phrases. Here we introduce a non-invasive decoder that reconstructs continuous language from cortical semantic representations recorded using functional magnetic resonance imaging (fMRI)...