The Lucas–Penrose argument uses Gödel's incompleteness theorems to claim human minds can do something no algorithm can — so we are not mere computers, and true machine "understanding" may be impossible.
The sketch
Gödel showed that any consistent formal system powerful enough for arithmetic contains a true statement it cannot prove ("this statement is unprovable here"). Lucas and Penrose argue: a human can see that the Gödel sentence is true, while the system itself cannot derive it — so human insight transcends any fixed algorithm.
If we can always "step outside" a formal system to see truths it can't reach, no algorithm fully captures the mind.
Why most AI researchers reject it
- A human reasoning is (arguably) a formal system, with our own unprovable Gödel sentence — we don't actually escape the trap.
- We can't even verify our own consistency, which the argument quietly assumes.
Still, it's one of the sharpest challenges to "the brain is just a computer," and a natural foil to The Chinese Room and the Turing Test.
Related: The Chinese Room · The Hard Problem of Consciousness · The Turing Test