Does AI have human-level intelligence? The evidence is clear
20 hours ago
- #General Intelligence
- #Turing Test
- #Artificial Intelligence
- Alan Turing's vision of human-level machine intelligence is now a reality, with AI systems like GPT-4.5 passing the Turing test 73% of the time.
- Large Language Models (LLMs) have demonstrated broad cognitive abilities, including solving complex mathematical problems, generating scientific hypotheses, and composing literature.
- Despite these achievements, 76% of leading AI researchers in a 2025 survey doubted that scaling current AI would yield Artificial General Intelligence (AGI).
- The debate around AGI is complicated by ambiguous definitions, emotional fears of displacement, and commercial interests distorting assessments.
- The authors argue that by reasonable standards, including Turing's own, current AI systems exhibit general intelligence comparable to humans.
- General intelligence does not require perfection, universality, human similarity, or superintelligence, but rather sufficient breadth and depth of cognitive abilities.
- Current LLMs meet criteria for general intelligence at both basic (Turing-test level) and expert levels, surpassing many human benchmarks.
- Common objections to AI's general intelligence, such as being 'stochastic parrots' or lacking world models, are addressed and refuted by the authors.
- The authors emphasize that intelligence does not require embodiment, autonomy, or agency, countering anthropocentric biases in the debate.
- Recognizing AI's general intelligence is crucial for policy, risk assessment, and understanding the nature of mind and the world.