Essential Math for AI

Chapter 14. Artificial Intelligence, Ethics, Mathematics, Law, and Policy

Torture the data enough and it will confess to anything.

Nobel Laureate and economist Ronald Coase (1910–2013)

AI ethics is a wide and deep topic, and it is emerging as a new area at the intersection of the philosophy and AI fields. We can only scratch the surface in this chapter, highlighting some issues and possible ways to address them, but leaving many equally important ones out. Nevertheless, this chapter has a message that I don’t want you to miss:

We need more of us to be situated in both AI and policy.

In my learning journey from math to its applications in AI, I discovered that AI should not be disentangled from policy, and that the two should evolve together. I can sit and write about the million examples where there are ethical considerations associated with AI technology, such as data security, privacy, surveillance, democracy, freedom of expression, workforce considerations, equity, fairness, bias, discrimination, inclusivity, transparency, regulation, and weaponized AI, but this is not how I will approach this subject. My take on these issues is from a slightly different angle, where I have seen firsthand how people try new weapons on populations in war-torn areas, and yet the governments and the media deny, do not comment, or say the unfortunate events were mistakes, that they will be investigated, then all move on to better things. When there is a new technology that affects people at scale, ...