The Leiden Declaration on Artificial Intelligence and mathematics https://leidendeclaration.ai/ is disappointing. The preamble is good; it clearly explains the threats that mathematics faces due to AI. The reason it is disappointing is that the only proposed solution is more transparency.

In the preamble, the authors state: « […] the consequences of large-scale investment in artificial intelligence are being widely discussed in regard to warfare, mass surveillance, political disruption, and environmental damage.». They are right: AI is currently being used to kill, surveil, and impoverish people. AI for mathematics cannot exist independently of the rest of the world; for it to function, companies need to make a profit, they need to sign deals with the military, and they need to keep destroying the planet to power their servers. To use AI is to contribute to this system. No use case of AI has provided sufficient benefits to justify the destruction it is causing.

To have as the only (possibly) constraining recommendation to « Evaluate the ethical consequences of your work, and take action accordingly» is simply not understanding the problem. AI can never be ethical in a capitalistic society. The purpose of this recommendation is to open the way to people claiming that they are using AI in an ethical way, which is impossible. How can one claim to be for open science and not condemn the use of AI, which is based on stolen work?

The few mathematical problems that will be solved by AI won’t outweigh the years of work and the brilliant minds we are losing because of it. No mathematical problem is worth the death of people, the destruction of our education system, and the destruction of our planet.