Turing Test 2.0: A Better Way to Test Machine Intelligence?
On this episode of Mind Matters News, host Robert J. Marks speaks with Dr. Georgios Mappouras about his proposal for a more rigorous test for measuring artificial intelligence. Mappouras argues that the original Turing Test is not enough to determine true intelligence in AI systems, as it focuses too much on simulating human-like conversation rather than demonstrating genuine understanding and problem-solving abilities. Mappouras proposes that a true test of general intelligence should involve an AI system being able to take in non-functional information — information it does not yet know how to use — and then demonstrating the ability to extract new knowledge and functionality from that information, similar to how a human can derive new insights and capabilities from observing the world around them. He believes current AI systems have not yet passed this higher bar for general intelligence, and he is skeptical that large language models like ChatGPT have truly achieved human-level intelligence, despite their impressive natural language abilities.
Additional Resources
- Georgios Mappouras “Turing Test 2.0: The General Intelligence Threshold”
- Robert J. Marks “Is AI Truly Creative? Here Is the Ultimate Test”
- Alan Turing “Computing Machinery and Intelligence”
- François Chollet “On the Measure of Intelligence”
- Robert J. Marks “Handbook of Fourier Analysis & Its Applications”
