Arguments Map
Building a comprehensive map of the arguments around AI safety and AI risk — their assumptions, their sources, and how claims relate to one another. The aim — make it easy for anyone to locate where they stand and what follows. Currently in Obsidian; will become its own site. This is one of the main activities I decided to pursue at AFFINE 2026.
Motivation
At this point in history, I believe it is utterly ridiculous to not consider AI as a threat to us all, especially given the extremely precarious international landscape. We need action and we need it worldwide. This work wants to lay down all the concerns surrounding AI safety and AI related risks, considering the evidence for each in a (as much as humanly possible) unbiased way, so that anyone can make an informed opinion on where they think this technology stands, which are the risks they care about, and where they can be impactful.
Structure
How nodes, edges, and assumptions are organized in the current Obsidian vault.
PRELIMINARY SURVEY:
Artificial Intelligence is the hot topic of the decade, probably of the century. Everybody is talking about it and confusion is reigning the communication channels. There are those who confidently claim that in a couple years (or five or ten) we will have super-intelligent systems, capable of surpassing humans in basically every aspect: someone thinks that such systems will be mainly beneficial to humans or at most neutral, an ultimate tool; someone else is concerned that such ASI (Artificial SuperIntelligence) will decide to kill all humans or anyway seize control. On the other side, someone smirks at them: in the end, we are stochastic parrots, right? and no real understanding occurs: they are just repeating patterns they found in data and they will never be able to actually pose a threat to humanity, neither be extremely useful — unless of course they get used by ill willed individuals. More (in my humble opinion) reasonable approaches, identify a technology — which is showing remarkable capabilities and which is extremely difficult to properly understand, interpret or control — being deployed at large scale in key infrastructures around the world: we do not know if these systems are going to act as we think they will, and yet we are withdrawing our agency and oversight in exchange for automated processes. You do not need to be a genius to see that this is (at least potentially) a huge risk for humanity. Finally, we already witness a large amount of concrete problems emerging from the use of artificial systems, from biases towards consolidated forms of power, psychological damages, job losses and uncertainty, mass surveillance, automated weapons and so on and so forth. People worry that the long-term threats are being used to actually shift attention from the short-term (already happening) ones, while also working as a marketing move (for all those companies that, more or less transparently, have the military on top of their customer list): extremely dangerous implies extremely powerful.
Roadmap
What it will look like once it becomes a public site.
How to contribute
Once public, how others can suggest claims, sources, or counter-arguments.