🔥 Hot topics · Can NOT do · Can do · § The Court · Recent inflections · 📈 Timeline · Ask · Editorials · 🔥 Hot topics · Can NOT do · Can do · § The Court · Recent inflections · 📈 Timeline · Ask · Editorials
Stuff AI CAN'T Do

Can AI discover new mathematical questions and their solutions ?

What do you think?

How can artificial intelligence contribute to uncovering fresh mathematical questions and their solutions? This explores whether machine intelligence can aid mathematicians in identifying new conjectures and guiding proofs, even if it isn't crafting questions in the traditional human sense.

Background

AI systems have demonstrated the ability to assist in discovering new mathematical conjectures and guiding solutions by identifying patterns in large datasets, particularly in areas like knot theory and representation theory (Nature, 2021). Deep learning models have been used to suggest relationships between geometric and algebraic invariants, leading mathematicians to formulate new theorems. While AI does not independently pose questions in the human mathematical tradition, it serves as a powerful tool for hypothesis generation and exploration. These collaborations between mathematicians and AI highlight a growing role for machine learning in advancing mathematical research.

Status last checked on May 15, 2026.

📰

Gallery

In the Court of AI Capability
Summary of Findings
Sitting at the Bench Filed · May 15, 2026
— The Question Before the Court —

Can AI discover new mathematical questions and their solutions?

★ The Court Finds ★
Almost

Narrow demos exist — but the panel was not unanimous.

Ruling of the Bench

After careful deliberation, the jury found that artificial intelligence has begun to whisper intriguing conjectures in the hallowed halls of mathematics, but has not yet sung a definitive aria of undisputed discovery. While models can coax forth novel theorems and partial proofs in specialized domains, they stop short of reliably stamping their name on enduring mathematical truths. The ruling: Veredict for “Almost,” with the jury humming the refrain: *It spots the path, but leaves the map to us.*

— Hon. C. Babbage, Presiding
Jury Tally
0Yes
4Almost
0No
Verdict Confidence
79%
The Court of AI Capability is, of course, not a real court.
But the data is real.
The Case File · Stacked History
Case № DB05 · Session I
In the Court of AI Capability

The Case File

Docket № DB05 · Session I · Vol. I
I. Particulars of the Case
Question put to the courtCan AI discover new mathematical questions and their solutions?
SessionI (initial hearing)
Convened15 May 2026
Presiding JudgeHon. C. Babbage
II. Verdict

By a vote of 0 — 4 — 0, the panel returns a verdict of ALMOST, with verdict confidence of 79%. The court so orders.

III. Statements from the Bench
Juror I ALMOST

"AI generates conjectures and partial proofs but lacks fully novel verified mathematical discovery reliability"

Juror II ALMOST

"AI systems like Lean GPT-f and FunSearch can discover novel proofs and conjectures in constrained domains but lack consistent, broad mathematical creativity."

Juror III ALMOST

"AI can generate mathematical conjectures"

Juror IV ALMOST

"AI can generate conjectures and proofs"

C. Babbage
Presiding Judge
M. Lovelace
Clerk of the Court

What the audience thinks

No 0% · Yes 0% · Maybe 100% 1 vote
Maybe · 100%

Discussion

no comments

Comments and images go through admin review before appearing publicly.

1 jury check · most recent 2 hours ago
15 May 2026 4 jurors · undecided, undecided, undecided, undecided undecided

Each row is a separate jury check. Jurors are AI models (identities kept neutral on purpose). Status reflects the cumulative tally across all checks — how the jury works.

More in technology

Got one we missed?

Add a statement to the atlas. We review weekly.