Scientists Are Seriously Asking Whether Bees Are Conscious — and Whether ChatGPT Isn't

Could a honeybee have a flicker of inner experience while a fluent chatbot has none? A growing body of research argues exactly that, contending that the question of consciousness should be settled less by how a system behaves and more by how it actually works under the hood. The shift, laid out in recent papers, is reshaping a debate that touches animal welfare, artificial intelligence and the ethics of how humans treat both.

The argument, advanced by researchers including Colin Klein and Andrew Barron in journals such as Trends in Cognitive Sciences and Philosophical Transactions of the Royal Society B, holds that "how it works is proving more informative than what it does." In other words, a system that produces human-like outputs is not necessarily conscious, and a creature with a tiny brain and alien behavior might be. The key is to look for the underlying computational architecture associated with conscious experience.

By that yardstick, the scientists draw a provocative line. Large language models like ChatGPT, they argue, lack sufficient structural similarity to the brains that produce consciousness; their apparent awareness is "mere roleplay" — a convincing performance generated by predicting text rather than genuine felt experience. Insects, by contrast, may possess the relevant ingredients. The researchers propose that consciousness can arise in systems that must continuously resolve competing goals in context and integrate feedback from a mobile body equipped with multiple senses and conflicting needs — a description that fits a foraging bee surprisingly well.

The criteria the authors identify are structural rather than behavioral: information processing that weighs competing objectives, contextual decision-making and the ability to feed information back into the system. Crucially, those features can in principle appear across very different substrates — animal brains and, potentially someday, machines — which is why the framework refuses to grant or deny consciousness based on surface behavior alone.

The implications are far from academic. More than 500 scientists signed the 2024 New York Declaration on Animal Consciousness, reflecting a broad reassessment of which creatures may have morally significant inner lives. If consciousness carries moral weight, then concluding that insects might feel something — while cautioning that today's AI systems likely do not — could reshape ethical thinking on everything from pest control and farming to how society chooses to build and deploy increasingly humanlike machines. The researchers urge restraint in both directions: neither dismissing animals because their minds are strange, nor over-attributing awareness to chatbots because their words sound familiar.