chapter five

5 Back to humans

 

This chapter covers

  • Illustrating why human cognition remains a central reference in AI debates.
  • Assessing how artificial neurons are inspired by the human brain.
  • Identifying the main initiatives in modelling the human brain.
  • Comparing how artificial neurons differ from human thinking.

For much of its history, artificial intelligence has advanced by borrowing ideas from the human mind while steadily distancing itself from the human brain. Artificial neural networks (ANNs) began as an attempt to capture something essential about biological learning, yet they matured into systems shaped more by mathematics, data, and computation than by physiology, retaining only what could be formalized and scaled. This tension lies at the center of contemporary debates about thinking machines. When artificial systems produce language, recognize images, or adapt to feedback, it is natural to ask whether they think in ways similar to humans, or whether the resemblance is only superficial.

5.1 Why it comes back to humans

5.1.1 Using ourselves as the reference

5.1.2 The language of minds

5.1.3 Why the metaphor endured

5.2 A source of inspiration

5.2.1 From observation to abstraction

5.2.2 Designing for trainability

5.2.3 Learning as weighted influence

5.2.4 Keeping representation diffuse

5.3 Modelling the human brain

5.3.1 Why model the brain at all?

5.3.2 Levels of biological fidelity

5.3.3 Focusing on the wiring

5.3.4 Efficiency in human hardware

5.3.5 What brain modelling offers

5.4 Differences with artificial neural networks

5.4.1 Learning without consequence

5.4.2 Intelligence without agency

5.4.3 Memory as structure, not state

5.4.4 Uniform time and frozen memory

5.4.5 Representation without understanding

5.4.6 Why surface similarity misleads

5.5 Summary