Research Benchmarks LLMs on Human-Like Understanding and Sensory Experience

A new benchmark study analyzes the capabilities of large language models in areas of human cognition. It specifically tests models from OpenAI and Qwen on their understanding of sensory experience and subjective consciousness, known as qualia. The research aims to explore the boundary between narrow AI, which excels at specific tasks, and a theoretical general AI possessing genuine understanding. It poses fundamental questions about what machines can compute versus what humans intuitively experience.