A recent research study by the University of Waterloo surveyed 300 Americans and found that two-thirds of the respondents (67%) believe that generative AI, especially ChatGPT, shows some form of consciousness.
In the study, respondents were asked to rate, from 1 to 100, whether they believed ChatGPT was conscious. Of the 300 respondents, 100 gave it a score of 1, and the rest of the respondents went higher than 1. However, the thing to note here is that two-thirds responded to varying degrees of consciousness.
This result shows that people are attributing human qualities to AI and over-believing its powers, which eventually has the potential to give rise to various AI ethics and legal issues.
There is also a folk psychology hidden behind this reasoning. People believe in the commonsense framework of mental states, and generative AI is similar to human cognition, thus, people automatically apply human perceptions to AI.
Although OpenAI has developed a lot in recent years, and it has even developed a new scale to measure how close its models are to achieving artificial general intelligence (AGI), even the most promising experts don’t think that AGI systems will be capable of true emotions.
There was also a correlation between how often a person uses AI and how likely they are to perceive consciousness. The truth is just that ChatGPT is good at mimicking humans, and the AI model doesn’t seem like a human brain at all.
The professor of psychology and co-lead of the study at the University of Waterloo, Dr. Clara Colombatto, explained. “These results demonstrate the power of language because a conversation alone can lead us to think that an agent that looks and works very differently from us can have a mind.”
The researchers are planning to look deeper into what is causing people to believe that AI has consciousness and what that means for an individual or a society.
