Tests for consciousness in humans and beyond

Color perception, robots, and safety

• Discussion opens with color perception: humans can’t easily “swap” red and blue, but robots could remap labels or latent spaces.
• Some argue robots can operate directly on spectra, with well-understood mappings; color naming is constrained, not arbitrary.
• Others note safety systems (traffic lights, signs) are redundant via shape/position, so an “AI virus” confusing colors seems less plausible than direct sabotage.

What is consciousness and when does it occur?

• Multiple states mentioned: development, disorders, brain injury, anesthesia, sleep, meditation, drugs.
• One view: “pure awareness” persists through sleep; others dispute this and attribute post‑sleep reports to bodily signals and memory.
• Phenomenal consciousness is often defined as “what it feels like” to be a system, even if content is minimal.

Definitions vs tests

• Many note that “consciousness” is overloaded: used for intelligence, language, subjective experience, etc.
• Some say the article sensibly targets phenomenal consciousness specifically, but others doubt any test is possible because consciousness is first‑person only.
• Counterpoint: operational tests (e.g., for anesthesia decisions) can sharpen the concept and are practically valuable.

Intelligence vs consciousness; animals and AI

• Strong emphasis that intelligence and consciousness are orthogonal; thermostats and calculators can be “smart” in narrow ways without being conscious.
• Several insist dogs and other mammals are clearly conscious; some argue public and scientific views on animal consciousness have shifted over time.
• Others allege “moving goalposts”: as AI gains capabilities, people retreat from intelligence-based criteria for consciousness.

LLMs, p‑zombies, and machine consciousness

• Views range from “LLMs already meet earlier criteria for sentience” to “LLMs are p‑zombies / fancy spreadsheets.”
• Disagreement over whether claims like “LLMs have no desires or pain” are justified without mechanistic models of these in humans.
• Comparisons made between chatbots, image models, “hello world” programs, and chairs to challenge where, if anywhere, we draw a consciousness line.

Philosophical stances and thought experiments

• Hard problem of consciousness, Mary’s Room, p‑zombies, solipsism, panpsychism, dualism, and idealism are all referenced.
• Some suggest consciousness may be emergent from complex sensory integration; others say it cannot be simulated or is non‑physical.
• A neurobiologically grounded theory (neuronal group selection, Darwin automata) is proposed by one commenter as a promising unifying framework and path to conscious machines.