There are suggestions that AI could revolutionize research in science and math in the coming years, but humans could still continue to have a vital role to play.
That’s the view of Terrence Tao, a UCLA mathematician widely regarded as one of the world’s leading mathematical minds. The Fields Medal winner, often called the “Mozart of Math,” recently offered a nuanced perspective on how artificial intelligence will reshape scientific research over the next two decades—one that resists both excessive hype and undue pessimism about AI’s capabilities.
“In the next 10, 20 years, humans and AI in that time period will have complementary strengths,” Tao explained. His vision centers on a collaborative model where each party—human and machine—contributes what it does best.
According to Tao, AI’s primary advantage lies in its extraordinary ability to process vast quantities of information. “AI is just extremely good at synthesizing very large amounts of data,” he noted. “A human cannot read a million different papers or whatever and try all the ideas in all those papers, but an AI can kind of do that now.”
However, Tao is quick to point out where humans maintain a decisive edge. “Humans right now can just see five or six examples of some math problem and just see, okay, I see the pattern now,” he said. “They can generalize from very, very small amounts of data, which AI cannot do right now. Or they can try to fake it, but it is very inefficient at it.”
There has been lots of progress in AI usage in math in the last few quarters. Harmonic, a mathematical AI startup founded by Robinhood CEO Vlad Tenev, has said that its reasoning model Aristotle has solved an Erdos Problem. Several other startups are also using Lean to use AI to solve math problems.
For businesses and research institutions, Tao’s framework suggests a pragmatic path forward: invest in AI systems that can handle data-intensive tasks while preserving and valuing human capacity for creative insight and pattern recognition from sparse information. The most successful organizations over the next decade may well be those that master this complementary division of labor.