When a honey bee dances to signal the location of flowers and ChatGPT generates a paragraph on the nature of existence, are they both conscious? That seemingly absurd question is now the subject of rigorous scientific inquiry, reshaping how researchers approach one of humanity's deepest mysteries.
For centuries, consciousness has been treated as distinctly human—a measure of our moral standing and a line between us and everything else. But the explosion of artificial intelligence and growing scientific evidence about animal cognition has forced a reckoning. In April 2024, forty scientists gathered in New York to propose something radical: the New York Declaration on Animal Consciousness, which has since been signed by over 500 scientists and philosophers. The declaration argues that consciousness is realistically possible not just in familiar mammals, but in all vertebrates—reptiles, amphibians, fishes—and many invertebrates including octopuses, crabs, and insects.
Simultaneously, the rise of large language models like ChatGPT raised an unsettling possibility: machines might be conscious too. Five years ago, philosopher Susan Schneider suggested that if an AI could convincingly discuss the nature of consciousness itself, it might well be conscious. By that standard, we would now be surrounded by conscious machines. Yet these debates have rested on surface appearances—how things act, not how they think.
Recent scientific papers propose a fundamentally different approach: looking at the machinery itself. A new paper in Trends in Cognitive Sciences, coauthored by Colin Klein, examines the structure of information processing in AI rather than its behavior. The researchers identified key indicators of consciousness based on how brains and computers actually process information—whether they can resolve trade-offs between competing goals, whether they contain informational feedback loops, whether they integrate information in specific ways. What matters for consciousness, the authors argue, is not what you do, but how you do it.
The verdict for current AI? Blunt and clear: no existing system, including ChatGPT, is conscious. The appearance of consciousness in large language models is achieved through a mechanism too different from biological consciousness to warrant calling it genuine. Yet the door remains open. Future AI systems with different architectures might cross that threshold.
For animals, scientists are taking a parallel path. Researchers have proposed a neural model for minimal consciousness in insects, published in Philosophical Transactions B. Rather than getting lost in anatomical detail, the model focuses on the core computations that simple brains perform—the ancient evolutionary problem of coordinating a mobile body with multiple senses and conflicting needs. If scientists can identify the specific computation that generates experience, they will have created a level playing field: a way to compare consciousness across humans, invertebrates, and computers using the same framework.
This shift matters profoundly. If consciousness isn't uniquely human, then our moral obligations expand. Philosopher Jonathan Birch calls this the precautionary principle for sentience: even when uncertain, we might err on the side of caution by assuming something could be conscious. A bee tending a wound, an octopus solving a puzzle, a future AI weighing competing goals—each might deserve consideration not as metaphor but as possibility. Science is no longer asking only whether other minds exist. It's asking how to recognize them.