Twenty-five mathematicians who have won the Fields Medal, the most prestigious award in their field, have signed an open letter warning that artificial intelligence poses a danger to mathematics. The letter was published on the Math and AI portal.
The signatories said the goals of companies that develop artificial intelligence and the goals of the mathematics community have seriously diverged. The letter states that the drive by AI companies to solve mathematical problems in order to demonstrate their own effectiveness is causing harm to the science.
The mathematicians said they see this misalignment as part of a much broader problem that also touches other scientific and creative professions, and society as a whole.
What the letter says
The authors of the letter noted that over recent months, news of neural networks succeeding at complex problems has been discussed well beyond the scientific community. But in their view, solving a problem is only a tool and a means toward the real goal, which is a deeper understanding of mathematical concepts.
They warned that a rapid, machine-generated solution leaves no time to identify new methods along the way, and creates problems around authorship and plagiarism. The mathematicians acknowledged that artificial intelligence can speed up scientific work, but said the consequences of using it will depend entirely on people and the principles they apply.
Why the Fields Medal carries weight
The Fields Medal is awarded once every four years to mathematicians for outstanding achievement, and is widely regarded as the highest honor in the field since there is no Nobel Prize in mathematics. A letter signed by 25 of its recipients represents a significant share of the small group of living laureates who have ever received the award.
A contrasting view on AI safety
The letter follows earlier comments from cybersecurity expert Alexander Dvoryansky, who said artificial intelligence today remains fully under human control and was built from the outset on a principle of doing no harm to people. Dvoryansky made the remarks while commenting on reports that American companies are developing emergency shutdown mechanisms for neural networks designed to prevent dangerous behavior by the technology.
