AI is making mathematical history. Experts explore how recent breakthroughs could reshape research, what people bring to the equation and why studying math still opens many doors.
A few short years ago, asking AI to add two four-digit numbers together for a third-grade math assignment would have been a bad idea. Over half the time, the answer would be wrong, studies have shown. Long division or a “leaky tank” problem, an algebra class staple? Forget about it.
Although AI still struggles with arithmetic, in 2024, Google DeepMind’s AI systems scored high enough for a silver medal in an elite high-school competition that tests mathematical reasoning rather than basic calculation. A year later, OpenAI grabbed the gold.
That was just a warm-up.
In May 2026, AI made headway on a question that mathematician Paul Erdős posed in 1946. Given a fixed number of dots on a flat surface, how many pairs can sit exactly one unit apart? Experts had a long-standing prediction, but the bot overturned it once and for all.
In a crowning moment in September, OpenAI announced that the Navier Stokes problem, which describes how fluids move, had been solved. The answer helps scientists understand when computer predictions of air or blood flow become unreliable.
The 88 hours it reportedly took might sound like a lot. Still, it’s nothing compared to the roughly 90 years people spent trying to put the question to rest.
These well-defined problems make for useful test cases, Northeastern mathematics professor Ken Duffy told Northeastern Global News.
So are mathematicians about to get taken out of the equation?
Experts in mathematics and computing put the breakthroughs in perspective. Here are five takeaways from the conversations.
1: There’s more to math than getting the answer in the back of the textbook. #
Schools and universities aim to equip future experts with the tools of the trade.
“We’re teaching people stuff that we already know to be true,” Duffy said.
However, the field amounts to much more than mastering the material. Research math is a different ballgame.
“The currency of research is (finding) a good question to ask,” Duffy said. The payoff might be new technology, or simply an elegant proof.
2: AI is taking a crack at the hard stuff. #
Within research math, AI excels at processes that call for brute force, such as gathering tons of data, researchers say. It also has a knack for linking facts together at breakneck speed.
To tackle the Erdős problem, AI applied techniques used to study numbers to geometry and found an unexpected connection specialists missed.
Solving the Navier-Stokes equation took around 10,000 bots sifting through decades worth of knowledge in hours, according to an announcement last month by OpenAI.
3: AI can solve a problem while still leaving people in the dark. #
Mathematicians earn recognition for explaining how they reached a solution and developing methods along the way, Duffy said. A good proof lays out the reasoning, gives others a route they can learn from and opens up fresh territory for exploration.
However, AI falls notoriously short at showing its work. The proofs it provides often leave mathematicians scratching their heads.
“When we mark students in exams, we don’t give them full points for getting the correct answer if they don’t explain how they got there,” Duffy said.
4: Still, AI is earning its keep well beyond math. #
AI solutions can be tested in Lean, a proof-checking system, he added. Some have been retracted after initial release when they didn’t pan out.
Drawbacks aside, the computing power of AI has transformed many fields. In biology, it predicted the shapes of over 200 million proteins, helping scientists determine their function.
Mapping cells in bone marrow to study changes induced by leukemia used to be a painstaking task, Duffy said. It called for staining tiny fragments with fluorescent proteins and taking loads of images.
“You’ve got 60 million or 100 million cells in there,” he said. “And then some poor graduate student goes blind looking at each image and drawing circles around (them).”
Duffy’s team farmed out the task to AI, sparing graduate students time — and eyesight.
The AI accurately picked out individual cells in the images, mapping the bone marrow with a fraction of the human effort, Duffy said.
James Boyd, co-founder of the Wolfram Institute, a research nonprofit dedicated to exploring the computational universe, told Northeastern Global News that AI is also challenging assumptions about math as a whole.
“It would be healthy for the field to broaden beyond proving theorems alone,” he said, suggesting that with AI in the picture, settling theoretical questions is no longer the main event.
One direction Boyd brought up is experimental mathematics. It focuses on drawing conclusions from patterns, much like a biologist tracking monarch butterflies or watching cells divide under a microscope. Computers, in turn, do the numerical heavy lifting and AI tests if the patterns hold up.
“I like to call this ‘high-performance human mathematics,’” Boyd said. It taps into the computational prowess of machines but “remains grounded in human methodology,” he added.
The key is keeping the step-by-step process of getting to the answer front and center. With humans in charge of those steps, they’re unlikely to get left by the wayside in favor of the results, he added.
5: A math degree still adds up to plenty of opportunities. #
Duffy once overheard a coffee shop client questioning a relative’s decision to study math. Wasn’t he going to end up as a schoolteacher?
The punchline: the guy landed a well-paid job at Capital One, one of the 10 largest banks in the U.S.
Teaching and higher education are honorable pursuits, but math majors have many other options, Duffy said. Finance and engineering companies gobble up graduates for their ability to crack tough puzzles.
Besides, the bots haven’t put critical thinking out of business as of yet.
“AI still hallucinates,” Duffy said. “It still produces gobbledygook.”