Abstract:
Fields Medal winner and famous mathematician Terence Tao recently issued a warning about mankind’s increasing reliance on artificial intelligence. He worries that as people gradually hand over their thinking and problem-solving processes to AI, humans are losing their understanding of the problems themselves and their solutions, and this trend may soon have serious consequences in the real world.
These remarks came from Terence Tao’s speech at the Science x AI Summit held at the University of California, Los Angeles on October 16. The Science and Artificial Intelligence Research Foundation (SAIR) recently released a complete speech video titled "Math 2.0" on YouTube, and the relevant clips subsequently attracted attention on the overseas social platform X.
"We are entrusting our thinking to AI and letting them solve problems for us. But at the same time, we are also losing a guarantee - we cannot be sure whether the solution given by AI is really what we want." Tao Zhexuan said, "I am really worried that we will soon see the consequences of this in the real world. This is actually a worry that I have outside the field of mathematics."
Tao Zhexuan pointed out that the progress of artificial intelligence in the field of mathematics is changing the traditional way of solving problems. In the past, in order to solve complex problems, mathematicians needed to study and understand relevant theories for a long time, and accumulate new knowledge in the process. But nowadays, AI can give answers quickly, but humans may not be able to understand the reasoning process behind the answers.
He believes that the risk of such changes is relatively controllable in the field of pure mathematics, because mathematical proofs can usually be verified through formal tools. But when the same approach is used in the real world, the situation becomes much more complicated.
Tao Zhexuan specifically mentioned the fields of power grid, air traffic control, drug research and development, and chip design. He worries that in the future, people may directly hand over complex problems to AI for processing, but lack a full understanding of the underlying model and real-world constraints.
In this case, the solution given by AI may be completely established in the mathematical model, but it does not meet the real needs of the real world. AI may even find loopholes or boundary conditions in the model, meeting superficial indicators while deviating from the goals humans originally hoped to achieve.
Tao Zhexuan emphasized that the problem does not lie in the AI technology itself, but in how humans use AI. He advocates promoting mathematical research into the so-called "Math 2.0" era, so that AI can not only help humans solve problems, but also enhance human understanding of problems and help researchers identify the limitations and potential risks of models.
On October 22, Terence Tao also published an article in Nature magazine, calling on the mathematical community to re-examine research goals in the AI era. He pointed out that mathematical research is facing new challenges as AI generates mathematical proofs faster than humans can understand and digest these results.
Tao Zhexuan believes that AI tools have the potential to enhance human understanding, but this requires changes in research methods and incentive mechanisms. What really needs to be retained in the AI era is not just the ability to obtain correct answers, but also the human ability to understand, test and judge answers.
Comments