https://www.nytimes.com/2023/03/27/science/ai-machine-learning-chatbots.html
Today, many modern predictions center around the idea of artificial intelligence. This article leads to the greater discussion of intuitive psychology, more widely known as theory of mind, which essentially connects humans to their social surroundings. However, when looking at it through the lens of machines, could machines also develop their own theory of mind like humans?
To delve deeper into the wider discussion of intuitive psychology, Michal Kosinski, a psychologist at the Stanford Graduate School of Business, proposed a new perspective where he claimed that indeed, "Machines have theory of mind" (Whang, 2023). Therefore, through Dr. Kosinski's claim, readers can draw the conclusion that there must be fewer uncertainties associated with a model that is large enough with lots of variations. Yet, his publication was quickly criticized by Tomer Ullman, a psychologist at Harvard University, who suggested that machines can contribute to greater uncertainties when there are slight adjustments or wordings that can greatly mess up the results. Therefore, with further investigations conducted by Maarten Sap, a computer scientist at Carnegie Mellon University, on the theory of mind tests, they prove that machines could not really comprehend and master the abstract and flexibility of human minds. Therefore, machine knowledge, described by these scientists, is "brittle" (Whang, 2023).
In both perspectives, the "Padua Rainbow," Prediction Framework, and the uncertainty discussions are all applicable in the discussion of AI and AI modeling. To further explain, both sides undertake the process of the Padua Rainbow in leading to their prediction of how machines can in the future reach the capacity of human minds, yet both reached two very different conclusions. Additionally, both discussions incorporated the predictive framework of #HUMAN and #NON-RANDOM as human components were utilized as a comparison to the development of non-random models that were continuously tested.
As a result, even in the mainstream media today, whether machines can overtake humans' theory of mind is still up for a heated debate, and according to a 2022 survey, among the researchers, almost a balancing amount of 50% supported Dr. Kosinski's claim while the other 50% opposed his idea.