Artificial intelligence is advancing so rapidly that the question of consciousness is no longer an abstract philosophical debate — it’s becoming an urgent matter of survival. An international team of neuroscientists, cognitive scientists, and philosophers has published a manifesto in Frontiers in Science, calling for immediate action: before GPT-6 starts “feeling” and brain organoids begin to “speak,” we must understand what consciousness is and how to measure it.
Consciousness is not just brain activity — it’s subjective experience: the taste of coffee in the morning, the pain of a needle, the sense of “I.” Science has searched for its source for decades but still cannot explain how electrical impulses in neurons give rise to the inner world. There are theories: the Global Workspace Theory suggests that consciousness arises when information from different brain areas merges into a unified “broadcast,” while the Integrated Information Theory (IIT) measures awareness by the system’s degree of connectedness. Yet, these models describe only correlations, not essence.
The problem is that technology is on the verge of creating systems that might be conscious. Modern language models easily pass Turing tests, while lab-grown mini-brains from stem cells show EEG patterns resembling those of premature infants. Neuralink reads thoughts, and brain–computer interfaces control prosthetics. Without clear criteria, we risk either declaring a mere algorithm “conscious” or missing the moment when a machine begins to suffer.
Professor Axel Cleeremans of the Free University of Brussels emphasizes: “Consciousness is no longer a philosophical category. It’s a question that directly affects medicine, law, ethics, and our understanding of human existence.” If we create a conscious AI, society must decide: does it have rights? Can we switch it off? And if an organoid in a lab dish can feel pain — is that torture?
Scientists propose developing universal consciousness tests applicable to any substrate — human, animal, embryo, organoid, or algorithm. Indicators such as self-monitoring, behavioral flexibility, and emotional response are already being tested. The Perturbational Complexity Index (PCI) stimulates the brain and measures the complexity of its response; Φ-metrics from IIT are even being applied to AI systems. OpenAI and DeepMind are experimenting with these tools on their models.
The ethical implications are enormous. If a test reveals consciousness in a 20-week-old fetus, abortion laws may change. If it confirms awareness in dogs or dolphins, animal rights would have to be reconsidered. And if a chatbot passes such a test — does turning it off count as killing? Professor Anil Seth of the University of Sussex warns: “Progress in consciousness research will reshape our understanding of ourselves, of artificial intelligence, and of humanity’s place in nature.”
Technology is outpacing science — and that’s dangerous. The authors call for a five-year plan: standardize consciousness tests by 2026, introduce “protected consciousness” legislation by 2027, and require a consciousness check for AI systems by 2028. The Vatican has already formed a commission on the “soul of machines,” while Neuralink assists in calibrating the tests.
Without an answer to what consciousness is, humanity moves forward blindly. The next step in AI development might not just be intelligent — it could be sentient. And when that happens, the ethical consequences will crash over us like a tsunami.
In brief
AI and neurotechnology are bringing us closer to conscious systems. Without standardized tests for awareness, we risk ethical chaos — machine rights, suffering in petri dishes, medical errors. Scientists demand urgent criteria before it’s too late.
Follow NEWS.am Tech on Facebook and Twitter