Full disclosure
Believe it or not, this starts with Minesweeper. Specifically a "Triple-A Minesweeper" that briefly took gamedev spheres by storm. Don't ask me to tell you what it's about, I did not check. Before I saw enough posts about the thing to move past apathy and click the link, it inevitably turned out that the dev had used generative AI tools.
It's not the first time something like this happens and it won't be the last time. We have an entire meme for this kind of scenario: we call it a milkshake duck, courtesy of a 2016 tweet by pixelatedboat.
The incident, minor as it is, prompted a couple interesting conversations I wanted to formalize in longer form.
Due diligence
It's easy to mock people who fall for it. Can't you tell? Weren't the signs there? Can't you do due diligence? At least delete the post when it turns out it doesn't match your principles, geez!
Well, yes. It'd be nice if more people did that. Click through to the profile first, verify what the account is about, see where and when and what the context is, try to assess good faith, learn the tells, think twice before sharing, reposts are not endorsements but check anyway... This is what I do myself. I have a finely honed bullshit-o-meter by now and basically treat anything I see online with maximum suspicion until I do at least a cursory vibe- and fact-check.
Turns out, that's a miserable way to live. It makes you into the killjoy, the party pooper, the one who says "And this is why we can't have nice things" again and again and again and again until people start resenting you. In some extreme cases, it will get you labeled as a bigot or a wrecker for refusing to look past inconvenient facts on your own side.
Deputizing everyone as an amateur fact-checker is not the path towards a better world. Verification is not an easy craft to practice, it does not scale. More importantly, it simply should not be on the audience to be inherently suspicious beyond basic media literacy. Defaulting to distrust is a form of hypervigilance and that's most famously associated with trauma.
You've no doubt seen testimonies to that effect: generative AI tools have ruined everybody's fun because you can no longer assume something nice is real. Is that cute animal real? Was that video faked? Can I guarantee that this silly game did not put money in the pocket of people who want me dead?
It forces us into paranoid readings, in the media analysis sense: we pick things apart searching for signs of bad faith. All too often, we'll find something, real or illusory. It's killing us.
The net effect is that of grinding constant pressure that either turns you into the aforementioned unpopular killjoy (hi), or gets you to break and stop fighting from sheer exhaustion. That's the whole point.
Disclosure
Instead, people who use generative AI tools to produce any part of a work should disclose said use, in a way that is fully visible on the work itself. Regardless of legislation, it's just the right thing to do. Unless you live in a cave without internet - in which case, how are you even using the things? - you have to be aware that a lot of people are upset about those technologies.
In one of those conversations I alluded to, someone asked why a developer would disclose if they do not consider their AI use to be interesting. This, in my opinion, misses the point entirely: it is not about what you find interesting, it is about having consideration for your audience and their principles.
Put another way: you, as a maker, are the one with easy access to information about how your work was created. Surfacing that information does not take a lot of effort on your part. Reverse-engineering the modes of production is a lot harder if not impossible.
Therefore, you should be putting the info in a place where it is easy to find. Not in the small print. Not in a follow-up post to the announcement which you know is going to get the most traction. There's no need to play coy - you know full well why you're not doing that, and it's because you don't want to get yelled at.
Don't get me wrong: I deeply resent how many people on the "anti-AI" side of things seem to consider harassment a viable way to achieve change. I do not enjoy conflict for conflict's sake and I fully believe that treating our enemies as fellow humans is vital to avoid corroding our own hearts.
And yet, there's only one side that decided to run roughshod over climate goals, intellectual property, consent, and worse. If you're marginalized on any axis, consider whether you've ever defended your side's toxicity because that's the result of being under extreme pressure or even existential threats. (Looking at you, trans pundits)
Being afraid of getting yelled at is a mostly predictable consequence of aligning yourself with planet-destroying fascists. Consider that you might be doing something morally wrong. Most importantly, remember that you can always stop.
At the very least, do us a kindness and disclose your AI use so we can choose whether to engage with your work. Think of it as listing the allergens in your home bake - which is, I assume, self-evident if you do not wish to kill your guests.
Doublethink, doublespeak, doubletongue
Between my first draft of this and actually sitting down to finish it, another quality example dropped in the form of Molleindustria admitting that Future? No thanks! included some AI-generated content and featured a tiny amount of AI-assisted code.
Molleindustria, if you're not familiar, is a mostly-one-person independent game development studio that has built a name on radical leftist politics and commentary on current issues. The entire goddamn point of the studio's existence is to prove that videogame development and leftist politics can and must go together. It has a manifesto, for fuck's sake.
Future? No thanks! was released on September 4th, 2026. On October 17th, 2026 - i.e. over a month later and well after the initial burst of launch sales will have happened - the developer put out a blog post with the cursory title of "Future? No Thanks! AI disclosure". It's a classic example of self-apologetic doublethink: the article starts by laying out that yes the author is keenly aware that so-called AI is very bad, only to admit to using it anyway.
Praxis for thee but not for me, I suppose?
Believe it or not but this wasn't even the worst part. The worst part was at the end and got edited out once someone with a large amount of followers excerpted it to ask (paraphrase) "What the fuck, dude?". It was restored when the author got yelled at some more. I'll grant the guy grace he has not earned and refrain from discussing it in detail.
There is now an addendum which is more pathetic still:
Edit: These extra notes made some people upset and turned me into a strawman for arguments and positions I abhor. They were beyond the scope of this disclosure, not useful, and not even well written. I’m leaving them here in case you are trying to figure out what the fuss was about. These are hot, complex issues that deserve a more extended and thoughtful treatment. I’ll let smarter people hash out the broader ethics and politics of generative AI.
Again, this is a studio founded on and known for radical political commentary. Instead the analysis here is generic, milquetoast and self-absolving: rather than resolve his own contradictions by admitting his mistakes and changing his behavior, the author pretends that morals were never clear-cut to begin with. It serves as an especially glaring example worth studying but sadly this exact behavior and apologia is becoming more and more common.
The behavior change did happen, in this case. That it only happened after getting yelled at and with phrasing so petty it is worthy of a 5-year-old child told they should stop throwing toy cars at their classmates' heads is a disgrace.
His audience, his students and his peers all deserve better. Grow a goddamn spine. Stand up for something. Otherwise, how are you different from the large studios with PR departments that you so love to mock?
Deeper dive
This is a quick write-up but I'm also working through the following related ideas.
Software supply chains are historically fully opaque to the end user, and only considered by developers when it comes to security. Needing to account for upstream ethics is fairly novel... or is it? Yup that one's going to have to be about the goddamn white supremacists again.
Said opacity has also long since been weaponized by programmers to shut down conversations. I hold instead that non-techies do in fact have the right to tell us how to do our jobs, because we live on the same fucking planet. We either start doing better or get used to being hated and scorned.
None of us are strangers to relying on ethically compromised products. Conflict minerals, exploitative labor, animal cruelty, etc. It doesn't mean industries should not be regulated via mandatory audits, environmental standards, traceability and labor organizing. Every link in the chain has a responsibility to do its best.
Much like enforced suspicion is a poisonous inversion of responsibility, being forced into a defensive position is dancing to our enemies' tune. This applies to rhetoric but also to how we build resilient cultures and resilient tech. How can we defend ourselves more smartly than by slapping Cloudflare in front of everything? IMO static websites that are considerably cheaper to serve are part of the answer.