Neuroscientist Proposes 'Scientific Discovery' Test for Smarter AI, Not Just Language Skills
• Neuroscientist David Eagleman proposes a new "test of intelligence" for AI involving scientific discovery rather than just language processing.
• Eagleman explains that while AI like ChatGPT can eloquently synthesize information, it doesn't actually understand what it's saying.
• Eagleman illustrates that absorbing vast information isn't the same as intelligence, using the "Chinese room thought experiment."
• Eagleman differentiates between "level 1" and "level 2" scientific discovery, the latter involving true conceptualization and creativity.
• Eagleman sees AI as potentially creative but not yet able to filter ideas in the way humans would care about.