AI model with the ability to read the human mind using similar technology to ChatGPT: All you need to know

The world is artificial intelligence is emerging with new possibilities every day. What seemed impossible yesterday is now turning into a reality. As per the study published in the journal Nature Neuroscience, AI models based on the similar technology of Open AI’s ChatGPT and Google’s Bard can now read a person’s thoughts and convert them into text. This study was conducted by a doctoral student in computer science Jerry Tang and an assistant professor of neuroscience and computer science, Alex Huth at the University of Texas at Austin. 

What was the study? 

As a part of the study, three subjects were asked to listen to stories and were assigned to MRI machines. After this experiment, scientists claim that they were able to generate the thoughts of the subjects without the help of any brain stimulation. The team then used a customized Chat GPT-like  AI model to identify this brain activity and translate it into text. However, the subjects' exact thoughts couldn't be captured, and only an idea of what the subjects were thinking was translated by AI. 

What were the results of the study? 

As AI was only able to identify the gist of subjects’ thoughts, the results were only accurate up to 82%. In terms of decoding perceived speech, the accuracy of the model was 72 to 82%. Whereas, in terms of decoding imagined speech, it was around 41 to 74% accurate and while interpreting silent movies, the accuracy was from 21 to 45%. 

Are there any concerns about this breakthrough? 

Although the new development in terms of an AI model with the ability to read the human mind is surely a leap forward in terms of technology. However, some concerns cannot be ignored. As per the scientists of UT Austin, this breakthrough will help mute people to express themselves. However, this feature has to go through a lot of work before it can be used in a clinical setting. 

The scientists are also concerned about maintaining the mental privacy of the subjects. “We take very seriously the concerns that it could be used for bad purposes and have worked to avoid that," said Jerry Tang. They also conveyed that the decoder requires the voluntary cooperation of human subjects to function properly. Lastly, they have warned that this technology can be misused for criminal purposes such as employer and government surveillance.