While people around the world still have different opinions in understanding and accepting the effects Artificial intelligence like chatbot With ChatGPT, scientists continue to use the capabilities of pre-trained generative AI models. According to Scientific American, one of the newest applications of GPT-based models is the direct interpretation of human thoughts.
A research team in collaboration with the University of Texas has developed a new GPT model, the details of which have been published in an article in the journal Nature. This model interprets people’s brain activity through blood flow. The technology, known as GPT-1, allows researchers to access what users hear or picture without the need for surgery or the use of complex devices and equipment.
Researchers note that due to the use of fMRI technique in GPT-1, this model cannot analyze some specific words that may be in people’s minds. Since the mentioned technology works at a higher level of abstraction, that is, instead of searching for meaning, it extracts and understands the meaning from people’s brain activity, some details are lost while analyzing the thoughts. For example, one participant in the study listened to a recording of him saying, “I don’t have a driver’s license yet,” but processing fMRI data generated from the moment he heard the words translated the original sentence into “he hasn’t started driving yet.” is” changed. Thus, GPT-1 cannot currently transcribe our thoughts word for word, but it understands their true meaning.
But the model of artificial intelligence interpreting human thoughts raises a very important question in the mind: where will this technology end?
In theory, technology is not inherently destructive. TomsHardware He writes, technology is a concept that can be used to achieve a goal. GPT-1 can help ALS patients communicate better with others. In addition, technologies like the mentioned model can be used to record people’s thoughts and open a new space for self-improvement.
But not all applications of GPT-1 are beneficial, and for example, it can be used to directly extract brain information from a subject who does not want to present his thoughts. On the other hand, the relatively easy way of using this technology is both useful and harmful.
In their article, researchers have mentioned the possibility of misuse of GPT-1 technology and have provided explanations about its negative effects. Now it seems that we have reached a stage where even our thoughts are no longer safe and artificial intelligence has acquired the ability to extract our thoughts.
Source link
