How can AI chatbots replace human doctors when they can’t even pass a dementia test? It never occurred to me someone would ask this question but here we are.
Comments
Log in with your Bluesky account to leave a comment
You can’t measure human cognitive impairment in LLMs because they aren’t human. It’s like testing a car for a kidney infection when it has a leaky brake line
Poor title: you can't show cognitive decline unless you were previously at a higher level to decline from. The title should be "chatbots now test at the same level as a human with minimal cognitive impairment." Pretty impressive if you ask me.
Comments