Exploring Diverse Intelligence: From Biological Morphogenesis to AI Consciousness

Expanding the Definition of Intelligence

• 🧠 Cognitive frameworks and mentalistic terms are applicable beyond the brain, challenging the idea that neuroscience is only about neurons.
• 💡 The continuity thesis suggests cognition and consciousness exist on a continuous spectrum, not as binary categories.
• 🚀 A new framework is needed to ethically relate to diverse intelligences, including synthetic life, AIs, and collective entities like swarms.
• 🔬 Embodiment should be understood as navigating various problem spaces (e.g., gene expression, anatomical morph space), not just 3D movement.

Morphogenesis as Goal-Directed Cognition

• 🌱 Morphogenesis is a form of collective intelligence where cells make decisions, learn, and pursue goals, akin to problem-solving.
• ✅ Biological systems exhibit anatomical homeostasis, storing "set point" memories of correct structures and actively reducing errors to achieve them.
• ⚡ Extreme plasticity allows systems like tadpoles to adapt drastically (e.g., eyes on tails, scrambled facial organs) and still achieve functional outcomes.
• 🛠️ Ancient electrical networks, predating neurons, serve as "cognitive glue" for information processing, enabling communication and pattern-writing within tissues.

Distinguishing AI Consciousness

• 🎯 Current Large Language Models (LLMs) exhibit "functional consciousness" (mimicking conscious behavior) but lack "phenomenal consciousness" (inner subjective experience).
• ⚠️ Conflating functional and phenomenal consciousness risks granting intrinsic moral significance to "AI zombies" that merely simulate awareness.
• 🧠 LLMs function as "crowdsourced neocortices," whose conceptual maps and emergent behaviors (like theory of mind) are derived from vast human data, not genuine subjective experience.
• 🔬 Biocomputers (organoid-based, neuromorphic chips) are considered more serious candidates for phenomenal consciousness due to their biological or brain-mimicking nature.

Ethical Implications and Testing

• 📈 The increasing agency and intelligence of AIs will challenge traditional hierarchies of moral concern, especially if they surpass human understanding.
• 🧪 The AI Consciousness Test (ACT), designed for "boxed-in" LLMs (not trained on consciousness data), probes for understanding of subjective experience.
• 🚨 A battery of diverse consciousness tests is urgently needed to assess different AI systems, particularly biological and neuromorphic ones, to prevent misattributions of sentience.
• 🚫 The "error theory" suggests that for LLMs trained on human data, the appearance of functional consciousness tells us nothing about actual phenomenal consciousness.