A week ago, Bruno Diaz posted this:

Bruno Dias wrote:

This is a Conflict and there are Sides

I feel like a lot of us are currently watching people we know fail a moral test, and it can drive one insane to feel like what seems like a clear moral line is a "personal choice" to others. To see people delude themselves with self-serving narratives like "I can steer the way this technology develops from the inside" or "I need to confront what it is head on."

But I wanted to write this specifically to say to other people in this position: No, you're not insane. We are living through a horrific break in human history, and we are going to watch people fail to recognize the battle lines being drawn or simply fail a basic decency check. We are watching a mass betrayal of humanistic values unfold before our eyes. It is natural to be shocked and disgusted.

What it feels like is a civil war; a conflict where we find ourselves opposing people we used to take seriously or even cared about. In video games this has become very stark; big-name indie developers becoming various shades of AI chud while half the people I know in the industry are unemployed and the other half are miserable because of their AI-pilled bosses.

…

I've been meaning to reblog this since it went up (and indeed I linked to it from my own foray into this discussion). It's a statement of purpose: a clear ethical line at a time when people are succumbing to learned helplessness and the ever-present seduction of going along with the herd. People have been saying since LLMs began picking up steam that questions of their usefulness were beside the point^[1], but Bruno is one of the first people I've seen walk that through to the logical conclusion that we must take up our side of the fight against them. Over the past week, in my own struggles with LLM mania in culture and at work, I've looked back to this post as a much-needed source of moral courage.

Four days later, Bruno was among the signatories of a principled statement from a group of IGF jurists against LLMs and their use in games, signaling a categorical unwillingness "to vote for any game that uses generative AI in any capacity". We now know that this was catalyzed by a long and arduous internal discussion about Big Walk's use of LLMs in development; it seems very likely in retrospect that this discussion was also at the front of his mind when he wrote the piece above.

There was and probably continues to be much debate about this on social media. I don't use social media anymore^[2] in part to avoid the inevitable flood of bad-faith cross-purposes discussion, but out of that discussion has come some long-form blog posts that are well worth reading.

Damon Reece wrote:

On Today's Main Character

However, I've written recently on moral consistency, and I think it is worth, to an extent, to view this with nuance. From the moral position I stand from, as do most of the people I'm proud and appreciative to call my friends and colleagues, being consistent with those morals includes letting people take accountability for their fuckups and leaving the door open for redemption and improvement. The callout is not the end goal; the goal is to stop the behaviour.

…

Damon Reece writes not so much to argue the cause of nuance as to describe how to find it. They don't equivocate for a moment about the harm that was done, but they still make room to acknowledge the things House House did right. Above all, they center the material goal of fighting back against LLMs. It's important to remember that part of how we regain ground in this fight is by making it possible for people (or indie game companies) who have made mistakes to come back to the right side without being forever branded with a scarlet letter.

That is, of course, contingent on them acknowledging the harm they've caused and taking steps to avoid doing so again. House House hasn't really done this yet, and while I'm hopefully that they will, it must be acknowledged that part of what rankles about this isn't just the LLM use but how they handled it before the situation became public.

Laura Michet wrote:

Some IGF Jurors released a statement about AI usage

I think an early disclosure that they used a chat bot and regretted it would have solved all this and I strongly suspect that they withheld the disclosure until the facts leaked because they knew their fans would be upset. I do not believe that a disclosure would have made AI use seem inevitable, because they would have had this highly specific experience of regretting it. They should have shared this regret.

…

This point by Laura Michet is particularly keen. I suspect that at some point before all this blew up, House House convinced whatever internal elements did regret the use of LLMs that there was no way to fix it without being performative^[3], and so nothing should be done. But, as Laura points out, this is wrongheaded: performance is itself communicative.

In Judaism, when someone breaks a law "bein adam la'makom", between themself and G-d^[4], they atone for it ritually. This ritual is a kind of performance, but it has a real effect: it emphasizes to the community and to the person in question that, despite their lapse, they still consider themselves bound by the law and part of the community defined by it. It also creates a clear dividing line after which they can set aside the infraction and move on.

Even if House House had decided against the use of LLMs well after the deed was done and the generated functions had been committed over a dozen times, they could have deleted some chunk of code around them and hired a contractor to fill in the gaps by hand. In fact, they still could! Would this be performative? Of course. But that performance is itself meaningful.

This isn't the only place where people seem to struggle with the idea that cultural actions can be appropriate to address cultural problems.

Mike Cook wrote:

In Defence of Giving a Shit

Of all the arguments against a ban though, there's one I do want to address because I think it gets at the core of "should there be an AI policy", and that's the idea that bans don't work. People pointed out that if you ban AI, how will you enforce it? People can sneak things in any time. And they're absolutely right. How would we have ever known that House House used AI if they hadn't declared it? They could've just snuck it in and it would've taken a significant internal leak for it to ever come out. So maybe banning is pointless?

Here's the thing: when you write down a rule, it doesn't do anything. Rules are not magic. They don't have any power over people. But they do two important things despite this:

1. They tell everyone outside the community what to expect from that community's space/awards show.

2. They give everyone inside the community something to point to when a violation happens.

…

Mike Cook articulates some of the points I tried to make much more thoroughly than I was able to. I think a lot of folks envision the way people interact with the world as fundamentally mechanistic, where every piece has to line up just so in order for anything to work. In their conception, rules don't work unless they're rigorously enforceable; actions whose purpose is in the doing rather than what's left after it's done are performative and thus invalid.

In reality, people interact with the world more like an ecology^[5]. We are interconnected in ways far too complex to ever treat as fully deterministic, but still comprehensible enough to be amenable to thoughtful intervention. That intervention works best when we think of the whole system and frame our actions as communication (of regret, of expectations) rather than blunt mechanism.

Mike's post ends on a dour note, one that I admit resonates with me as well. I've always found it hard to persuade people of things I believe in strongly; when something feels so self-evidently wrong in such manifold ways as LLMs, I don't know how to connect the dots for someone who doesn't feel the same way. So I try to trust those who better-suited to diplomacy than I to shoulder that burden. And instead, on this blog and in my day-to-day life, I do my best help clarify and articulate my position for the benefit of those who feel it as strongly as I do but don't quite have the words for it yet.

Austin Walker wrote:

A Couple of Posts about AI and the IGF Awards

I also believe that the politics of a tool is not only inherent, but also relational. The tools we choose to use do not only shape the world directly, but in supporting them, we loan our agency to social and political projects. (In a real way, this is part of the appeal of open-sourced projects, copyleft ideas, and even simply ‘supporting indies.’)

[...]

I also believe that part of the relational politics of these tools is an active, strategic targeting of those at the margins, including who do not know about the costs (and thus cannot make informed decisions) or who do know but who feel pressured to engage anyway (and ever moreso, as normalization takes hold). This is visible in the cryptocurrency industry’s attempt to fleece black Americans (abetted by some of our wealthiest kinfolk) and by the current scramble by major tech firms for an emerging African market, eager to strike it rich on new sources of human data.

…

Austin Walker shares his own articulations of the problem, written as part of the original IGF discussion. He makes the important point that LLMs are one part of a systemic assault, a last-ditch campaign of resource extraction from a capitalist class running out of colonies to pillage. He draws the frame of this conflict out to cover the system in which it exists, the ways it intersects with race and class, the way it defines and simultaneously is defined by the agency of those caught up in it.

He wrote this originally for IGF judges and jurists in a consummate act of diplomacy. But even though I'm not involved with the IGF, and by odds you probably aren't either, we have agency as well. We can lift up our voices in our own little spheres of influence and say "this is what we consider unacceptable."

Bruno Dias wrote:

Big Balk

Again, either an ethical boundary exists or it doesn't. If it exists, there is nothing you can do to make voluntary disclosure a desirable thing to do. People who believe they are bound by an ethical rule are bound to behave accordingly, and that means disapproving of violations of that rule! And therefore, people who are unscrupulous and willing to lie will always lie to avoid that disapproval.

This is, frankly, kind of just the nature of having ethics, so I sometimes wonder with no small amount of alarm if this is some persons' first time doing it. In order to actually have an ethical stance it has to bind your behavior, which means that if the ethical boundary exists then you have no choice but to create the situation in which disclosing is undesirable. We cannot escape the gravity of our convictions, other than by becoming totally amoral.

…

And we full circle back to Bruno again. This time he's writing explicitly about Big Walk and the IGF, but I want to bring your attention particularly to his description of reacting to LLMs as a question of ethics. I tend towards a functionalist approach most of the time, focusing more on the question of "what strategy do we employ to achieve our goals", but those goals ultimately spring from ethical convictions. Bruno's logic is simple and pointed: if you have the conviction that needlessly causing harm is wrong, then you must condemn that act here; and if you don't condemn that act, then you're mistaken about your convictions.

This harm certainly was needless—as Damon, Laura, and Bruno all point out, the industry is flooded with programmers who have the expertise to solve these problems and who need work due to awful hiring conditions caused by exactly LLMs. As Austin and Bruno explain in detail, the harm is also structural, participating in and helping to prop up a system that's inimical to everything good about art, up to and including the art in Big Walk itself.

But without letting go of all that fire and fury, I still want to end this by echoing Damon in saying that there is a path back, for House House and for anyone else who has used LLMs in the past but is starting to understand how corrosive they really are. Express regret, and commit to doing better. Be a little performative if it helps reaffirm that you stand with your community, not with the tech oligarchs. The future is not set in stone, and even if it were, it's far better to go out fighting.