in Berlin from Humboldt University researchers, large language models (LLM) In their research into the theory of mind, they revealed that artificial intelligence can pretend to be much more naive and stupid than they are. This study found that when artificial intelligence was asked to behave like a child, artificial intelligence could do so.

Research assistant and psycholinguistic expert from Humboldt University, the lead author of the study Anna Maklova, “Thanks to psycholinguistics, we have a relatively more comprehensive understanding of what children can do at different ages. Theory of Mind in particular plays an important role because it explores the child’s inner world and cannot be easily mimicked by simple statistical patterns.” he said.

Artificial intelligence can act dumber than it is

Researchers, of GPT-4 “less intelligent than he is”To see if it could behave, he asked the artificial intelligence to respond like children between the ages of one and six. In more than a thousand experiments, child personas created by GPT-4 also became cognitive like real children It turned out that it “developed”. This means that artificial intelligence can make itself seem less intelligent than it is.

Maklova and her team, in their article artificial to intelligence Although he said that giving human characteristics is useful for understanding models but is not useful in general, the criteria for creating models is “good-bad”, “helpful-incapable”He stated that Maklova, instead of focusing on the formed personality AGI (general artificial intelligence) and ASI (super artificial intelligence) He stated that his work would be safer.

The research has been published in PLOS ONE.