I’ve been doing a bunch of Wikipedia editing recently, which has been nice. It’s pleasant to feel like you’re in a constructive community, and one of the ways that this community is built is through the assumption of good faith.
As one of the evident challenges to the project at large is how to deal with contributions assisted by (or entirely generated by) LLM, it has been on my mind that the LLM somewhat undermines this principle. Assuming good faith is obviously necessary for a massively collaborative project like Wikipedia, but it’s historically backed up by recognizing that a person making a misguided effort is still making an effort. That effort on the other’s part is skin in the game, and a demonstration that they at least think the project is important even if they have a different idea on what it should be doing.
That the LLM reduces friction and augments human effort is somewhat core to their product design at the moment. Advertisements entice the would-be user with the idea that the LLM can take over the fiddly busy work that they don’t want to do. Tracking down sources is exactly the kind of time-intensive task that many people would prefer farm out, and so of course Wikipedia is awash with folks who want something on the wiki changed, but not badly enough to actually bother doing the work themselves. When such a large proportion of the interactions with strangers come in the form of canned wheedling slop, it is hard to retain respect for the unknown other. They don’t care enough to write it themselves, why should I care enough to carefully consider a response?
I think of course that the humans who want to remain in a project that values humans have no choice but to try their best to assume good faith anyway, and that’s a worthy thing to strive for. But it is a shame to me that this optimizing drive has its sights set on removing the “friction” of human interaction. I attended a sort of town hall event last night in which Loyola Professor Joe Vukov presented and took Q&A on Magnifica Humanitas. One of the points that he focused on repeatedly was the argument from the Church and from ethical scholars that human relationships and interactions are uniquely valuable in a way that automation cannot replace. I hope that these arguments are heard and heeded. As it stands there’s a massive amount of investment behind the idea that human interaction is an inefficiency to be automated away. A world where we interact with others mostly through their autonomous “agents” will have lost the valuable experience of recognizing humanity in a stranger.