We need to be sure not to mistake AI intelligence for actual consciousness, researchers say
- Bookmark
- CommentsGo to comments
We are in danger of making a fundamental mistake about artificial intelligence, a group of researchers have warned.
A new paper from neuroscientists at the Université de Montréal and Johns Hopkins University says that we need to be sure to distinguish between AI’s intelligence and it actually being conscious.
Because AI chatbots are given to appearing impressively fluent and empathetic, they can make us believe they actually have a consciousness that can understand our words. But they argue that even if a system behaves intelligently and seems to respond convincingly about our emotions, that does not actually mean it understands them, cares or has any kind of inner experience.
The distinction is becoming increasingly important as chatbots such as ChatGPT and Claude become more advanced and people rely on them more for personal and emotional issues, they note.
In the new paper, they point to something called blindsight. That refers to a phenomenon where people receive damage to their primary visual cortex and are able to see nothing in part of their visual field, but can still guess at what is happening.
That shows that someone might be able to process information but that does not actually show they are having conscious experience of it.
“A person with blindsight can respond accurately to visual information without the conscious experience of seeing it,” said Vanessa Hadid, a postdoctoral researcher in psychology at UdeM and at the McGill University Health Centre. Dr Hadid is one of three co-authors on the paper, along with UdeM psychology professor Karim Jerbi, a researcher at Mila - Quebec Artificial Intelligence Institute; and John W. Krakauer, director of the Center for Restorative Neurotechnologies at Johns Hopkins.
It is not clear whether that kind of conscious experience will ever be possible in a computer system. But at the moment there is no indication that it is – and we need to be careful not to be tricked, the authors warn.
“Current AI systems do not feel anything and do not have conscious experience," said Professor Jerbi. "But the more fluently they speak and the more sensitive they seem to our emotions, the easier it becomes to forget that."
The work, ‘The illusion of AI consciousness: Lessons from human unconscious processing’, is published on the neuroscience site The Transmitter.
Join our commenting forum #
Join thought-provoking conversations, follow other Independent readers and see their replies
Comments