The world of artificial intelligence is brimming with possibilities. Interest in the AI space is at its peak these days with the popularity of chatbots like ChatGPT, new Bing and Google’s Bard. While artificial intelligence is not really a new concept (we’ve been using it all these years in some way or another), people today are certainly more aware and intrigued. And it won’t be wrong to say that we are standing at the edge of a technological revolution, of sorts. Now, Scientists developed new AI model that can convert human thoughts to texts.
Things that seemed impossible a couple of months back are seemingly becoming a reality. And the ability to decode human thoughts using technology is one of them. As per latest reports, scientists from University of Texas at Austin have made it possible to convert human thoughts to text. Led by computer science doctoral student Jerry Tang and assistant professor of neuroscience and computer science Alex Huth, the study is an important breakthrough in the AI world.
AI converts human thoughts into text
The scientists used functional magnetic resonance imaging (fMRI) machine to record 16 hours of brain activity from three human subjects as they listened to narrative stories. The researchers were able to identify neural stimuli that corresponded to individual words. The team then employed a custom-trained GPT AI model, something similar to ChatGPT, to decode this brain activity and translate it into text. However, the participants’ exact thoughts couldn’t be captured and only a gist of what the participants were thinking was translated by AI.
The scientists say that the results had an accuracy of up to 82 percent. In decoding perceived speech, the AI model was 72-82 percent accurate while in decoding imagined speech, the accuracy was around 41-74 percent. In interpretations of silent movies, the accuracy ranged from 21-45 percent.
The results were published in the Nature Neuroscience journal. What’s most fascinating is that this was done without the help of any brain implants.
“For a noninvasive method, this is a real leap forward compared to what’s been done before, which is typically single words or short sentences. We’re getting the model to decode continuous language for extended periods of time with complicated ideas,” Huth was quoted in a report published on the UT Texas website.
Concerns around the development
The scientists at University of Texas say that with this breakthrough tech, it will be possible to help the people who are physically incapable to express themselves. However, the tech is still developing and a lot of work will be needed before it can actually be used at a clinical setting.
On the other hand, scientists behind this tech say that their decoder requires the voluntary cooperation of human subjects to function. They also said that mental privacy of people should be respected when developing brain-computer interfaces.
Even though they developed new AI model, scientists have also expressed concerns over its possible misuse. They warned that the technology could potentially be used for nefarious purposes, such as government or employer surveillance. Concerns around mental privacy have also surfaced.
Original source: indiatoday.in
