AI outperforms humans in all mathematical research areas by 2028?
💡 What the odds say
The market puts this at about a 23% chance — unlikely.
No money — just record your call and see if you were right. Yes is at 23% right now.
The market heavily doubts AI will surpass humans across all mathematical research areas by 2028, largely due to recent high-profile demonstrations where AI models failed rigorous math tests that humans aced, shifting the burden of proof onto AI proponents.
📊 Base rate: Major scientific milestones like proving a mathematical theorem are rare; for instance, only a handful of AI systems have autonomously solved open problems in pure math, making a complete sweep across all areas within two years historically improbable.
What's driving it
- • A June 12 Nature article reported humans outperforming AI on a highly rigorous mathematics test, directly challenging the 'Yes' case (Nature, Jun 12).
- • On June 15, a separate test showed four AI models scoring zero on a math test they could not cheat on, while humans averaged a perfect 10, reinforcing skepticism (WION, Jun 15).
- • A theoretical article on July 2 discussed key AI vs human differences, potentially solidifying doubts about AI's current reasoning limits (Simplilearn.com, Jul 2).
The case for YES
- • Rapid advancement in AI deep research, as highlighted by comparisons of Claude, ChatGPT, and Grok on complex tasks, suggests AI tools are progressively closing gaps (AIMultiple, Jun 22).
- • Forecasts from May 2026 analyzing nearly 10,000 predictions indicate many experts see a path to AGI or superhuman reasoning within a few years, aligning with the 2028 deadline (AIMultiple, May 22).
- • AI's ability to specialize—like solving specific mathematical problems—could rapidly expand to cover all areas through iterative learning and scaling, even if general tests remain difficult.
The case for NO
- • The June 12 Nature test required rigorous proof and logic, areas where humans still demonstrably outperform AI, showing a fundamental gap not yet closed (Nature, Jun 12).
- • Four distinct AI models failing a math test without cheating (June 15) signals that current AI lacks robust, generalizable mathematical reasoning, a prerequisite for 'all areas' (WION, Jun 15).
- • The 2028 timeframe is short; even with rapid progress, achieving across-the-board superiority—from number theory to topology—within two years appears unrealistic given recent failures.
What to watch
- • Release of a new benchmark or test by a major research institution (e.g., similar to the Nature test) in Q3 2026 showing AI improvement would shift odds toward 'Yes'.
- • A public demonstration by a leading AI lab (e.g., OpenAI, DeepMind) of AI autonomously solving an unsolved mathematical problem by mid-2027 would boost 'Yes' odds.
- • Another failure of AI on a prominent math exam, like the one from June 12, or a human-exclusive competition win in late 2026 would solidify 'No' odds further.
AI-generated · grounded in recent news + odds · informational only, not advice. Verify on the source platform.
Data from Manifold’s public API, for informational purposes only. PredictPal is not affiliated with any platform and does not facilitate trading.
Discussion
Loading…
How it resolves
Resolved by whoever created the market, at their discretion per the question's description. It's play-money (Mana) and not tied to an official source — treat it as a community forecast.
ⓘ A market settles under its own written rules, which can lag what looks decided in the news — so the price may not move to 100% the moment an outcome seems obvious.
View the official rules on Manifold ↗Related markets
When will Starship flight 14 happen?
The field is extremely front-loaded: the top candidate (before 2027-04-01) commands 95% odds, and all ten candidates above 5% stack into the next nine months, reflecting a market consensus that Flight 14 is imminent but not immediate. The largest recent shift is the collapse of near-term dates after the July 17 abort, when the 'before 2026-10-01' candidate dropped from ~80% to 71% following a second launch abort.
11 outcomes
Apple Announces AI Glasses by September 30, 2026
2 outcomes
Will Anthropic release its next Mythos-class model to the public by August 31, 2026?
Yes ≈ 46% chance
Which of these Language Models will beat me at chess?
The field is extremely concentrated on 'any model announced before 2034' at 88%, but the long tail of 22 candidates above 5% suggests bettors see many plausible paths to a 1900-rated human being beaten by a future LLM, with the single biggest recent shift being the 2025-08-15 Business Insider report that OpenAI's o3 swept a chess tournament against xAI's Grok 4, likely boosting confidence in near-term AI chess ability.
52 outcomes