That’s part of what brought me to this conversation in the first place. When I am ready for release 6 months to a year down the road and you see Standstill Digital Media pop up in the forums, you will already know what I am about, regardless of whatever box your AI tag wants to put me in.
By this logic, I should also have the freedom to decide if I want to use addons that have errors in the console and don’t work at all, no matter what I do to it. If an AI tag is being forced on me, why isn’t a “this addon doesn’t work” tag not being forced on others?
As far as MIT is concerned, what do you think is going on with AI? Do you think it is writing entire libraries for me? Because it’s not. Mostly, it is just helping me think things through or suggesting ways MY code could be better. Git commits and documentation. Troubleshooting. Busy work kind of stuff.
I have to say, I tend to agree with you, more so after this conversation. Here’s an idea for all the AI doubters out there. Actually use AI on a project. And not some dinky 3rd grade science project. Something meaningful that would have an impact on the community. Something that is tested and works. And then come back here and tell everyone how much of that AI did on its own and how much of it came straight out of your brain.
It is 5:00 am here. I’ve got to get some sleep. Just try to see things from my perspective. Some of us aren’t doing what you think we are doing.
One thing I keep forgetting to mention. Earlier, I said Claude is my junior software engineer and I am the senior. That was a metaphor. Standstill Digital Media is just Claude and I. I’m not a professional software developer in any sense of the word. I’m just a hobbyist programmer with no formal training but I have been learning from books, online docs, and YouTube and whatnot since 1996 so I do know a thing or two about game dev.
Does it work? If so, it was mostly created and wholly tested by a human. (AI can do some cool stuff, but it absolutely can not create things at will and it can not test your code.) If it doesn’t work, does it matter who created it? That seems just as simple to confirm as “this addon doesn’t work” to me.
I can see I’m not going to win this argument. People will continue to operate under the preconceived notions they have about AI without actually understanding anything about its capabilities or impact on coding. That’s fine. They have every right to remain ignorant and I will go to my grave defending that right. But I do hope some of you will check out my stuff later, when it is ready, even though by then, it might be forced into the AI box that everyone is so eager to wrap things up in.
I think we’re talking about two different things here. My post isn’t focused on whether an addon works or not. I’m not saying an addon made by AI can’t work, or one made by a human can’t work. That doesn’t really matter. What matters is the freedom people should have in deciding if they want to have ANY AI generated / stolen code in their project or not. By saying “I will not tag my resource as AI assisted”, you are essentially telling people they have no right to know what license the code has that they’re about to include in their product. And even if they don’t care about licensing issues, they might still wish to decide to not use AI code anywhere.
By not tagging, you are essentially removing the community’s freedom to choose what they want to use and what they don’t. It’s not a “box” as you said, it’s basic transparency.
At what point did I say I would not tag my resource as AI assisted? I started this conversation by saying Claude handles my git commits. It signs every commit message with “Co-Authored by Claude.” It also shows up as a contributor on all of my repositories. Do I strike you as the kind of person who is trying to avoid transparency? Would I even be involved in this conversation at all if I was just going to ignore the tag anyway?
You really don’t understand what AI does at all, do you? It’s not stealing code. It’s not generating code. It is just helping me make MY code better. It may have been trained in some small part on stolen code. The same could be said for me and probably most human programmers. But I assure you, there is no stolen code in my repositories. It is 95% me with a little AI enhancement around the edges.
When I first started working with AI, I was under the same impression you are. That it is a magical wish box that you type a few sentences and ta-da! All your wishes are granted! That’s not it at all. I work hard. Day and night. It’s 6:00 am here and I should be in bed but I’m here, working. Working on my brand and the perception of it. I’m not getting paid for any of this. All of my repositories are licensed under CC0. I’ve even removed the copyright and trademark notice so everyone can use it in any manner they wish. I can legally do that because it is MY code.
I’m not saying I am refusing to be put in the box you are crying for me to be put in. I play by the rules and I don’t have any problem with transparency. I’m just saying that your little label idea isn’t going to accomplish anything positive. All the people who are writing AI slop are going to ignore your label and the people who actually want to be up front about what they are doing are going to be ignored by people who could actually benefit from what is being produced.
I appreciate you clarifying that you include a disclaimer in your git commits, this is good for all parties involved. But my point isn’t about your personal work ethics or how you decided to create your addons. Most people, especially those who are relatively new to game development, will not always check github commits, and won’t even know what those are. Sure, people might ignore the tag, and that’s a problem, but it’s not just to catch “bad” people who break rules, but, as I mentioned before, to give people the ability to decide for themselves if they want to use AI code in their projects or not. It doesn’t matter what reason they have for wanting to avoid it. It doesn’t matter how “little” AI code there is in your project.
If someone simultaneously insists on not using AI code and not doing any research whatsoever (in this case checking git) on tools they find online, it seems like either one or the other of these things is not very important to them. “This is very important to me but i dont want to spend an additional minute on researching it”.
Best solution: mark things “no AI” if it is important to you and your audience. Most people dont care and will not even participate in these discussions or be aware of these tags. If someone really wants to avoid AI, its the best way.
This conversation is starting to remind me of my first release when someone asked how could they know if my addon had viruses or other harmful code in it. I suggested they look at the code. It’s all freely available in plain text on Github. They responded that it was a lot of code to go through, line by line. Sorry, there is nothing more I can do for you. You either care enough about your project to inspect the code you are injecting into it or you don’t. It was not, is not, and will never be my responsibility to grade my code by your yardstick. Either do the work or don’t. Your choice will show in your product, not mine.
If you already include a disclaimer on your github repo, it shouldn’t be that much of an issue to include one on the forums. We’re not asking all AI submissions to go away. We simply ask a small tag to be included.
The moderators will decide what the best course for this should be. I’m obviously not the one who will implement this, this is the Forum Feedback category. I gave feedback, we’ll see how it’s implemented, if implemented at all.
My issue isn’t with transparency. I have already laid that out in as plain of English as I am capable. My problem is that your tag doesn’t actually solve anything. All it does is add another requirement for me to follow when I, just like anyone else like me, is already being as transparent as possible.
Do you think that hardened ex-cons pause even for a microsecond when they see a “Gun Free Zone” sign? I’m guessing you do.
Let the code tell the story of how it was created. Anyone worth being a part of this discussion should be able to recognize AI slop at a glance. And if you’re not willing to inspect the code you are injecting into your project, AI generated code is the least of your worries.
There is no current sane way to use those “AI” tools without taking into account their socio economic cost to all of us. That 20 dollars a month is really much more than that, and we haven’t included the exploitation of many people doing enshittified piece work in labelling all that data for ML and LLMs. You don’t give a shit, cool, but don’t come and ad hominem folks who disagree with you as “religious” with all the prejudices it implies.
In fact, one could say the religious stance is people still pushing AI at all costs in the face of reality. But I won’t
And ripped off code is why a lot of companies wouldn’t touch Open Source for a long time as they couldn’t or wouldn’t trust the vetting processes to include contributions in FOSS.
Cheers, keep on making cool games!
80% of games programming in not making the game its spent sorting out the bugs you get along the way, if you use ai and dont understand the code how you going to fix it, most of the time its not bugs its silly stuff like opps its not ment to go off the screen but thats because you have not told it to not to do that, i think they would spend more time prompting the ai tryng to fix things and intoducing more bugs then just programming. If you undersand your code you know exactly whats going on and where to fix it.
I didn’t describe people disagreeing with me as anything at all. I described the reasoning of some users as more religious than practical in nature when it comes to this specific topic. Standstill seemed to struggle to understand the point of view of most others here in the thread and only thought about the problem in practical terms. I rephrase it from “religious in nature” to “philosophical in nature”.
I still think that the only way for users who really wants to avoid any AI code/assets/sounds/etc whatsoever in their games would be to rely more on a “made without AI”-tag. Why? Because those users are the ones that care about it and search for it. You don’t find bread marked “Not gluten free” and gluten intolerant people studying the packaging for breads without any markings. There aren’t any "Not Vegan”, “Not Halal”, “Not Kosher”, “Not Fair Trade” etc markings either.
It just makes it easier for everyone if the people that care about something that most people don’t care about sort the tiny details out themselves. It will make it easier for yourselves to find what you are searching for too.
Another issue that arises today is that you might unknowingly use second hand AI. So much is already out there that isn’t clearly marked as “AI” that you just can’t know. There is to my knowledge no tool that would clearly discover the use of AI in something you’d like to use in your project.
Actually it’s event become a problem for AI companies them self, that don’t like to feed their AIs with AI generated stuff.
An additional aspect is copyright. It looks like, at least in US, that all AI generated assets will not have copyright.
When you label your work as in parts AI generated, than you have no ownership on these elements. It has legal consequences if you label parts as AI. Maybe in future you have to label your product (by law) if AI was used.
I think it’s funny, that all big studios that use genAI in production will lose the copyright/ownership (in parts) on their products. So they have to keep their workers, just to ensure the copyright.
Sure but then you are doing a Junior Software developer out of a job, a junior that will never actually be able to become a senior because there is no work for them to do.
Thats fine for an independent like yourself or most people on here, but when the likes of EA do it you end up with stuff like this.
