Neuroscientists from the Massachusetts Institute taught AI to predict human feelings.
Neuroscientists from the Massachusetts Institute of Technology (MIT) have created a computer model using artificial intelligence that is able to determine and predict people’s emotions. Information about the development was published in the British scientific journal Philosophical Transactions of the Royal Society A.
It is noted that the program with AI is able to calculate such human feelings as joy, embarrassment, confusion, regret and gratitude.
Rebecca Sachs, a professor of cognitive neuroscience at MIT, pointed out that they took the script of the British TV game Golden Balls to build the model.
In it, two participants are allocated $ 100 thousand After discussion, each participant must decide whether to divide the amount or play only for themselves. If both agree to divide the money, then each receives $50,100, and if only one of the players decides to play for himself, he gets all $<>,<>.
“From scant descriptions of events, observers can predict what emotions the people involved will experience. “Our model uses reverse planning to infer a person’s beliefs and preferences, including their social preferences,” the study said.
The model combines a person’s reactions to events to calculate how a person will eventually evaluate their other actions or what is happening around them before any visual reaction is manifested on their part.
“In the course of the work, neuroscientists taught the model to compare 20 emotions, including joy, relief, guilt and envy. The framework combines reverse planning, event evaluation, and emotion concepts into a single computational model,” the paper says.
In the summer of 2022, Google engineer Blake Lemoyne gave an interview to The Washington Post, in which he stated that the LaMDA neural network language model created at Google has its own consciousness. He also argued that AI was capable of feelings like a person.
The developer was initially suspended due to a violation of the company’s privacy policy. The company later fired Lemoyne. In May, he gave an interview to Futurism, in which he talked about Google’s attempts to take control of its version of AI, unlike OpenAI, which released a chatbot with ChatGPT AI with almost no control