“The real risk with artificial intelligence is not that it will become smarter than us, but that we will rely on it without understanding its limits.” - Gary Marcus, AI researcher.
Understanding why AI struggles with simple reasoning
Artificial intelligence may have mastered complex pattern recognition and optimization, but it still falters at tasks that require everyday reasoning. Neuroscientist and AI entrepreneur Max Bennett explores this paradox in his book A Brief History of Intelligence: Why the Evolution of the Brain Holds the Key to the Future of AI. His earlier book on the same theme is recognised as a masterpiece already. He now suggests that to make AI truly intelligent, we must understand how biological intelligence evolved.
Evolution’s five breakthroughs toward intelligence
Bennett identifies five evolutionary leaps that shaped intelligence: steering, reinforcement learning, simulating, mentalising, and language. These breakthroughs transformed organisms from simple jellyfish to humans capable of abstract thought. AI, he argues, has so far mastered only the first two navigation and trial-and-error learning, while still struggling with the last three, which involve internal modelling, empathy, and symbolic communication.
The missing leap from learning to simulation
Current AI relies heavily on reinforcement learning and neural networks, which predict patterns but cannot simulate the world. Humans, by contrast, constantly run mental simulations, imagining outcomes before acting. This capability stems from evolutionary structures like the neocortex and hippocampus, which maintain an internal model of reality.
Why scaling up is not the answer
Adding more neurons or larger datasets cannot solve AI’s common-sense gap. Bennett emphasizes that biological intelligence evolved through qualitative shifts, not mere scale. For AI to gain intuition, it must transition from prediction to simulation, developing a true understanding of context, not just statistical correlation.
The next frontier of intelligent design
AI’s future lies in replicating how the brain continuously compares real-time sensory data with its internal world model. Until systems can reason about causes, effects, and intentions, they will remain powerful yet fundamentally limited tools.
Summary
Artificial intelligence has advanced in pattern recognition but lacks the common sense humans evolved through simulation and abstract reasoning. Bridging this gap requires qualitative innovation, not just scaling up neural networks, by modeling how brains predict, simulate, and reason about the world.
Food for thought
Can AI ever achieve genuine understanding without first replicating the biological essence of human thought?
AI concept to learn: simulation learning
Simulation learning allows AI systems to build internal models of the world and predict possible future outcomes before taking action. It represents the shift from reactive intelligence to proactive understanding, a crucial step toward true common sense in machines.
[The Billion Hopes Research Team shares the latest AI updates for learning and awareness. This is not a professional, financial, personal or medical advice. Please consult domain experts before making decisions. Feedback welcome!]
COMMENTS