AI Helps Unlock Discoveries in Kurt Gödel Notebooks

The Austrian logician Kurt Gödel, widely regarded as one of the most original thinkers of the twentieth century, left behind tens of thousands of pages of unpublished work. His two incompleteness theorems, published in 1931, transformed mathematics and paved the way for the modern digital age. Since 2017, Finnish logician Jan von Plato has led a project at Helsinki University to study sixteen notebooks written by Gödel in the early 1940s. Containing 1,009 pages of densely written text about mathematics and logic, the material was largely composed in an obsolete German shorthand known as Gabelsberger.

Progress in deciphering the notes was exceptionally slow, averaging about one page per week, until von Plato enlisted the help of the AI model Claude. Within a couple of days, the artificial intelligence provided a comprehensive digest of the entire material. The analysis revealed that Gödel had proved hundreds of theorems previously unknown to the academic community, proving he was truly ahead of his time in his mathematical formulations.

Unlocking Lost Proofs

Von Plato explained that the workbooks contained dozens of anticipations of results that did not officially emerge until the 1970s and 1980s, alongside fourteen unexplored themes in logic. When von Plato tasked Claude with investigating whether Gödel possessed a proof for the independence of the axiom of choice, the AI indicated he was merely a single step away from the correct definition. The reconstructed proof was subsequently verified using the Coq proof assistant program.

Further analysis by Claude extracted 663 theorems from the workbooks. When these results were tested through verification programs, roughly four hundred were confirmed as stated, about two hundred required minor repairs for precision, and fifty were false. Researchers note that continued study of these manuscripts may uncover additional treasures across logic, physics, and cosmology.

AI analysis of Kurt Gödel's notebooks revealed hundreds of previously unknown mathematical theorems.

Why it matters

This discovery sheds new light on the capabilities of historical scientific manuscripts and demonstrates how modern artificial intelligence can accelerate academic research by connecting thousands of pages of complex data as a single piece. The findings offer historians of science a deeper understanding of one of modern logic's most influential figures.