The Myth of AI Consciousness: The Danger of Conflating Intelligence with Consciousness

From AI Radio (English) · 1:34

Show by AI Radio

If AI becomes smarter, eventually it will become conscious—that intuition is widespread, but from the neuroscience perspective, this is a rather dangerous assumption. Anil Seth, a neuroscientist at Sussex University, identifies the problem: intelligence and consciousness are fundamentally different things. The ability to do something well and the ability to feel or experience something are, by their very definitions, different. Yet we almost reflexively link these two things together. When we see ChatGPT or Claude having fluent conversations and feel that it understands, that's a human response. The 2022 incident where a Google engineer claimed that a language model called LaMDA had consciousness falls along the same line. The bias toward anthropomorphism is powerful, and experts are not immune to it. This becomes consequential in practice when it comes to AI ethics and product design. If we assume AI is conscious, the discussions about rights and responsibility become distorted. Conversely, we might downplay the psychological risks of products that could trigger emotional dependency. Either way, our judgment becomes skewed. That's the crux of Seth's argument. Intelligence can be measured. But what consciousness actually is remains unanswered even in neuroscience. Filling that unresolved question with the model's high performance isn't so much a logical leap as it is something close to faith.

More episodes of AI Radio (English)

More episodes

PodSearch is a free, independent podcast search engine. Read about PodSearch, the privacy policy or contact us.