The rapid development of artificial intelligence is creating significant progress in the field of mathematics. However, along with that is a wave of concern from academia about the potential risks when AI is used more and more widely without appropriate control rules.
Recently, hundreds of mathematicians signed the "Laden Declaration on Artificial Intelligence and Mathematics", a document built over 8 months by a group of 16 researchers after a conference held at Leiden University (Netherlands). The Declaration also received the support of the International Mathematical Union.
This move appeared just weeks after OpenAI and Google DeepMind announced results showing that their AI model is capable of solving problems that have existed for decades.
In which, OpenAI published a deduction model that found the solution to a famous geometry problem posed by mathematician Paul Erdős in 1946. Meanwhile, Google DeepMind claims that the AlphaProof Nexus system has solved 9 of the 353 Erdős problems that have not been solved.
However, mathematicians believe that these achievements need to be assessed cautiously when AI can create arguments that sound reasonable but are actually inaccurate, making scientific verification more difficult.
This increases the burden on critics and threatens traditional standards of accuracy, transparency and independent verification.
The "Laden Declaration on Artificial Intelligence and Mathematics" also raises many other concerns such as the fact that AI does not fully cite the origin of research works, the risk of creating inequality between researchers who have and do not have access to technology, as well as the trend of publishing results through press releases or online posts instead of official scientific review channels.
In addition, researchers also warn that the increasing involvement of technology companies in mathematical research may affect the autonomy of this field.
The "Laden Declaration on Artificial Intelligence and Mathematics" does not oppose AI, but calls for building principles of using technology in a transparent and responsible manner.
Recommendations include publicizing the use of AI in research, ensuring full recognition of human authors' contributions, building appropriate publishing guidelines and maintaining ethical standards in cooperation between academia and the technology industry.
