Terence Tao calls for a ‘Maths 2.0’ built around understanding, not just AI answers
In a new Nature essay, Terence Tao argues that fast AI-generated proofs make verification, explanation and attribution central problems for mathematics.
When producing a proof becomes easier than understanding it
In a Nature essay published on October 22, mathematician Terence Tao argues that AI is changing mathematics faster than its institutions can adapt. His central concern is not that machines can now assist with difficult problems, but that a rapid flow of AI-generated results can outstrip the time people need to check, explain and absorb them.
From proof scarcity to ‘proof indigestion’
Tao calls for a ‘Maths 2.0’ response to a world in which proofs may no longer be scarce. In his account, the valuable work does not end when a system produces a plausible argument. Mathematicians still need to determine whether its claims are valid, understand why its methods work, identify prior ideas accurately and decide which results are worth developing. He warns that a result can be technically available without yet being useful knowledge for the field.
The essay also connects those questions to practical applications. Tao notes that mathematical models increasingly influence decisions beyond pure research, including credit and financial-risk systems. If models or their reasoning become too complex to audit, errors can affect people even when an output looks convincing. That is an argument for stronger scrutiny and documentation, not a claim that every AI-assisted result is unreliable.
A viewpoint, not a performance study
This is an expert commentary, not a peer-reviewed evaluation of a specific AI system or a new mathematical result. It offers a useful lens for reading the fast-moving news around automated theorem proving: ask what was checked, what is explained, whose earlier work is acknowledged, and whether another researcher can inspect the route to the conclusion.