An update on the OFCOM situation: As you know, censorship around the world has been ramping up at an alarming pace. OFCOM, the UK’s communications regulator, has singled out our community, demanding compliance with their Online Safety Act despite our minimal UK presence. This is a blatant overreach, and they have been sending letters pressuring us to comply with their censorship agenda.
Our platform is already blocked by many UK ISPs, yet they continue their attempts to stifle free speech. Standing up to this kind of regulatory overreach requires lots of resources to maintain our infrastructure and fight back against these unjust demands. If you value our community and want to support us during this time, we would greatly appreciate any and all donations.
i'm not going to argue mass AI theft with a lawyer that thinks there is no theft. with billions of dollars from that theft, they have plenty of money and lawyers fighting the many copyright lawsuits.
the internet has turned to shit as it were already. hard to trust most anything, filled with bots, hucksters, shills, ads ads ads, affiliate links, misinformation, disinformation and overall enshitiffication.
i.e. you have "ads that look like legit Reddit user posts" on reddit. shills incorporated. reddit was built on users and communities, and now that's being sold off to AI companies. A handful of people will make lots of money thanks to hundreds of millions of users. Typical silicon valley, like the gig economy, which is basically modern day serfdom.
Reddit is garbage. My frustration comes from people not being able to stick to their guns. Too many are just okay with things getting worse and won't push back. It's enraging. But those people sign up to use platforms for free and then continue using them when the bad policies are put in place, with all the information they need available to know that their use of the platform is making the rich richer. I hate it, but the human behavior is a separate subject imo. I don't think people have a right to now say that their reddit memes and bad puns have value and the companies 'stole' it from them.
The posts that were flooding my notifications that bothered me were completely low-effort AI slop. One, as I said, was literally copy pasted dangerous misinformation about "mental health coaches" with literal easily debunkable information. It even was formatted like the AI. The title with the bullet points in that exact pattern, if you've used ChatGPT you know the one.
I say, don't copy-paste AI slop to start posts. It's fine for you to use it to format your original thoughts, or to word things better, or to give you a jumping-off point as to information (DO NOT USE AI AS A SOURCE. I've had that shit tell me to drink bleach, that Harry Styles was asian, and that mildly blocked tear ducts turn your skin yellow. None of these are exaggerations. If you use the AI, ask the AI for the exact source and DOUBLE-TRIPLE CHECK).
I am also neurodivergent, and while I don't use AI here out of principle, both my mom and best friend benefit from it as autistic people in the world. They also follow common internet sense.
Last edited:
Reactions:
EvisceratedJester, avoid, ForgottenAgain and 2 others
derpyderpins
Pollyanna, loon, believer in love, believer in you
The posts that were flooding my notifications that bothered me were completely low-effort AI slop. One, as I said, was literally copy pasted dangerous misinformation about "mental health coaches" with literal easily debunkable information. It even was formatted like the AI. The title with the bullet points in that exact pattern, if you've used ChatGPT you know the one.
I say, don't copy-paste AI slop to start posts. It's fine for you to use it to format your original thoughts, or to word things better, or to give you a jumping-off point as to information (DO NOT USE AI AS A SOURCE. I've had that shit tell me to drink bleach, that Harry Styles was asian, and that mildly blocked tear ducts turn your skin yellow. None of these are exaggerations. If you use the AI, ask the AI for the exact source and DOUBLE-TRIPLE CHECK).
I am also neurodivergent, and while I don't use AI here out of principle, both my mom and best friend benefit from it as autistic people in the world. They also follow common internet sense.
So because we have an issue with low effort posts we should... make it worse with more, easier to reproduce, versions of them? Already six feet under, might as well dig to the center of the earth, I guess.
I personally would rather 50 percent of my Recovery section notifications NOT be AI posts. At least low-effort human posts took enough effort to put their fingers on the keyboard and type. Plus, this stuff is extremely dangerous because AI can make harmful things sound really appealing to those struggling.
I would like those in Recovery to recover, not take bad advice and end up CTB. I would also like those in the Suicide Section to not end up horribly disabled and in pain from an attempt.
Reactions:
EvisceratedJester, ForgottenAgain, Whale_bones and 1 other person
derpyderpins
Pollyanna, loon, believer in love, believer in you
So because we have an issue with low effort posts we should... make it worse with more, easier to reproduce, versions of them? Already six feet under, might as well dig to the center of the earth, I guess.
Not at all. My point is we allow low effort and "inaccurate" (although no matter what there will always be disagreement on what's accurate) so maybe tighten those rules rather than trying to deduce the tool used to create the post, determine the percentage of the post that relied on said tool, and remove based on those calculations rather than the amount of "low effort" or wrong it is.
As I get all these responses, I also notice there are two distinct, and arguably conflicting, trends in the arguments.
1) ai should be banned because the way in which it makes the outcome is evil or tainted in some way; or
2) ai should be banned because of the quality of the posts that use it.
Right now these two groups of thought are happy to act as one, but (2) will continue to be less of a problem as ai develops and people get better at using it.
Considering that, banning ai posts for group (2) reasons seems foolish because the reasoning behind the rule will inevitably disintegrate. When it comes to banning content, it's always harder to remove a restriction, because people with the ability to control other's content tend to like exercising that control. You'll then have a ban on something that leads to good content because of the outdated opinion that it leads to bad content
Complaints about "generic" posts and/or low effort content in this day and age are funny to me. If it's giving wrong information, call it out/correct it like you would a non ai post, considering people get shit wrong all the time.
Finding proper sources and reading them, explaining why something is false and trying to phrase it in a concise/understandable way, all of this takes time and effort. People who want accurate information shared shouldn't have to use their time and energy to fact-check and correct posts that weren't even written by another human. It's the poster's responsibility to stand behind the words they say, not everyone else's responsibility to make up for their lack of effort.
If people want to talk to an AI, then they just talk to an AI. Not go on a forum so other people can copy and paste what an AI pooped out.
As I get all these responses, I also notice there are two distinct, and arguably conflicting, trends in the arguments.
1) ai should be banned because the way in which it makes the outcome is evil or tainted in some way; or
2) ai should be banned because of the quality of the posts that use it.
The posters that have put forth their reasons for rules around AI usage here are all arguing for common-sense limits. No one is saying "AI should be banned". Tons of people have pointed out the ways in which it can be used beneficially and reasonably.
Reactions:
burningleaf, EvisceratedJester, avoid and 4 others
derpyderpins
Pollyanna, loon, believer in love, believer in you
Finding proper sources and reading them, explaining why something is false and trying to phrase it in a concise/understandable way, all of this takes time and effort. People who want accurate information shared shouldn't have to use their time and energy to fact-check and correct posts that weren't even written by another human. It's the poster's responsibility to stand behind the words they say, not everyone else's responsibility to make up for their lack of effort.
Again, a general principle I agree with. If my assistant drafts a document for me, I of course have to read it over before signing because it's my name and if it's wrong I face the consequences. Same if I use an AI. Same if I Google something and trust that source. Same if I use a source from social media. It's on me to verify before posting. So... What's different here?
The posters that have put forth their reasons for rules around AI usage here are all arguing for common-sense limits. No one is saying "AI should be banned". Tons of people have pointed out the ways in which it can be used beneficially and reasonably.
"Common sense limits." It's something subjective. An ultimate arbiter of truth will determine if you've used an appropriate amount of the tool or not, based on how it sounds to them and of course nothing else because there is nothing else. I've been around long enough to know that your of thing inevitably leads to whatever the thing is that's just being "limited" at first being fully banned.
"Common sense limits." It's something subjective. An ultimate arbiter of truth will determine if you've used an appropriate amount of the tool or not, based on how it sounds to them and of course nothing else because there is nothing else. I've been around long enough to know that your of thing inevitably leads to whatever the thing is that's just being "limited" at first being fully banned.
This thread feels like we're all putting our brains in a collective blender. I'll make it simple:
Should AI be allowed?Yes!
What are the requirements for this?That the AI be used either for ability/accessibility reasons, AND/OR properly sourced/double-checked for dangerous misinformation.
^ I feel like nearly everyone here can agree on this.
AI builds off of what it "absorbs." You remember that example I gave about it saying "mild tear duct blockage can cause yellow skin" right? When I checked the sources, it had confused the fact that there can be YELLOW DISCHARGE and RED/SWOLLEN SKIN.
Yellow skin is jaundice. Often caused by organ failure. Pretty fuckin bad. AI is not only not infallible, but unreliable most of the time!
Reactions:
Whale_bones, astr4 and Namelesa
derpyderpins
Pollyanna, loon, believer in love, believer in you
This thread feels like we're all putting our brains in a collective blender. I'll make it simple:
Should AI be allowed?Yes!
What are the requirements for this?That the AI be used either for ability/accessibility reasons, AND/OR properly sourced/double-checked for dangerous misinformation.
^ I feel like nearly everyone here can agree on this.
Plenty here have stated reasoning that doesn't agree with that, because AI is inherently evil from how it's made and/or because your aren't talking with a "human" even if it's accurate and human approved.
Allowing it for exceptions like ability/accessibility shows just how impractical the limits are. You see a post: now you have to determine - from the outside looking at only the end product - not only how much of it was AI, but what the person's reason was for using AI. It's not feasible.
AI builds off of what it "absorbs." You remember that example I gave about it saying "mild tear duct blockage can cause yellow skin" right? When I checked the sources, it had confused the fact that there can be YELLOW DISCHARGE and RED/SWOLLEN SKIN.
Yellow skin is jaundice. Often caused by organ failure. Pretty fuckin bad. AI is not only not infallible, but unreliable most of the time!
So someone posts something stupid and wrong (which happens all the time with or without ai): HOW are you determining that they got the wrong answer from AI? "It seems like it?" "Oh, you can tell"?
I'm repeating myself but can't remember who I've said what to, but this argument is also saying that as AI becomes more reliable and people get more familiar with how to use it properly the issue solves itself, so I don't see the logic in putting a rule in place that is guaranteed to become outdated sooner rather than later.
It's also important to note that this line of thought, if one accepts it as the valid reason for banning certain ai use, is inconsistent with the ai is evil/built on theft line of thought because this premise can be fixed.
Yes I'm weird, yes I'm stubborn, yes I'm annoying. But also yes all of you are wrong lol.
Generative AI is "evil" because it's based in stealing other people's content (yes, it's stealing, even if it's not illegal: law and moral are not the same), because it requires a lot of energy to do a job that it's mostly unnecessary and because it's destroying the life of a lot of people who are losing their jobs. Besides, the content that it creates is of poor quality and it's full of mistakes and misinformation, both accidental, because it lacks common sense and it can't understand what it's saying, and intentional, because it's controlled by corporations and governments.
Our world is horrible already, we don't need new problems.
In the case of this forum, I think it should be really restricted. Maybe there are some cases where it can be considered as a necessary assistant, but I don't really know any case like that. For now, I'm okay with the stance of the moderation team.
...if it's wrong I face the consequences. Same if I use an AI. Same if I Google something and trust that source. Same if I use a source from social media. It's on me to verify before posting. So... What's different here?
Having a post removed is a consequence of using AI as a copy-and-paste substitute for any human input. Instead of leaving AI completely unchecked and without limits, mods are forming guidelines so it's not left up to other forum members to use up their time and energy on interacting with copy-and-pasted posts.
"Common sense limits." It's something subjective. An ultimate arbiter of truth will determine if you've used an appropriate amount of the tool or not, based on how it sounds to them and of course nothing else because there is nothing else. I've been around long enough to know that your of thing inevitably leads to whatever the thing is that's just being "limited" at first being fully banned.
If your argument is with the amount of decision-making power mods have, that's a different subject. The instance that started this thread was one where the poster used AI to an extent that they didn't know if they'd written their own post. We have no reason to speculate about things like total banning when that doesn't align with what's happening on this site.
The large majority of posters here just want reasonable limits so the forum isn't filled with copy-and-pasted posts. The mod response in this thread says they've so far taken action when it's clearly such a case, and that they recognize gray areas and will continue taking into account how AI can be used for assistance. I think that's reasonable.
Reactions:
EvisceratedJester, NormallyNeurotic, astr4 and 3 others
derpyderpins
Pollyanna, loon, believer in love, believer in you
Having a post removed is a consequence of using AI as a copy-and-paste substitute for any human input. Instead of leaving AI completely unchecked and without limits, mods are forming guidelines so it's not left up to other forum members to use up their time and energy on interacting with copy-and-pasted posts.
You have no way whatsoever to tell when something is copy pasted and when it's not. And even if I'm wrong about that (I'm not) you soon will not be able to tell as the tech improves. You might as well ban anything typed on a white keyboard. You have 0 way to prove it, so the rule is meaningless.
You have no way without asking the person to tell which posts are copy pasted. And a good amount of people disagree, they want it banned entirely because AI is evil to them.
And again copy paste will continue to get more impossible to recognize as the tech grows and people get better at prompting.
So far I've been pitched: no ai at all; no """misinformation"""; no "copy paste"; some sort of percentage threshold of how much you used it; and probably something I'm forgetting. I remain right where I started: none of those make sense or are practical, and absolutely no one has an explanation for how to implement their rule other than it being "obvious" or "clear."
i think it's ok if they label it as assisted with ai. and give their reasons like English is not their first language or problems writing. i assume some give chatgpt their written summary and chatgpt makes a larger well written essay.
My schoolteachers taught me never to use Wikipedia as a source, and I believe AI falls into the same category. Just as you can reference the sources listed at the bottom of a Wikipedia page, you can use the source material from an AI-generated text, but not the Wikipedia article or AI-generated text itself, no matter how well-written or accurate the information may be.
Readers will naturally assume that you wrote the text yourself if you leave out the source. I think this is the main issue when someone posts AI-generated content without quotation marks and a source. Simply select the text, click the quotation mark button, and cite your source—no more complicated than the example below. If you skip this step, you are either trying to take credit for something you didn't write or you don't care much about the accuracy of the information.
Written by ChatGPT:
You should use quotation marks when copying text from another website word-for-word to indicate that it is a direct quote. Additionally, you should provide proper attribution by mentioning the source, such as the author's name and the website. This helps avoid plagiarism and ensures that you are respecting copyright laws and ethical writing standards. If you are paraphrasing, quotation marks are not needed, but you should still cite the source.
Adding quotes and a source to your AI-generated text without writing any additional text yourself may still be considered low-effort, but at least readers will know the text may contain false information because it was generated by AI. I think a post with credit where credit is due, whether it's to a user, web article or gAI, is far more permissible on forums than without.
On another note, using generative-AI to translate or rewrite texts in English doesn't seem to trigger detection tool that much. I wrote the first 2 paragraphs of this post in Dutch and asked ChatGPT to rewrite it in English with correct spelling and grammar—I wonder if you, the reader, had noticed. Most generative-AI detection tools concluded the text was likely written by a human.
All free generative-AI detection tool websites that Google listed on the first 2 pages. Results are for the first 2 paragraphs of this post, which is translated by ChatGPT.
You have no way whatsoever to tell when something is copy pasted and when it's not. And even if I'm wrong about that (I'm not) you soon will not be able to tell as the tech improves. You might as well ban anything typed on a white keyboard. You have 0 way to prove it, so the rule is meaningless.
There are plenty of websites that help you detect AI-generated text: Google results. Schoolteachers often use such tools to verify written work of students. AI detection software is not foolproof and gives you a probability percentage, so there's no definitive proof someone used generative-AI. But if a user shows a pattern of posting texts with a high generative-AI detection rate (low effort) then I think it's fair to remove those posts if a moderator also considers the content to be of low quality.
Last edited:
Reactions:
Whale_bones, pthnrdnojvsc and Namelesa
derpyderpins
Pollyanna, loon, believer in love, believer in you
My schoolteachers taught me never to use Wikipedia as a source, and I believe AI falls into the same category. Just as you can reference the sources listed at the bottom of a Wikipedia page, you can use the source material from an AI-generated text, but not the Wikipedia article or AI-generated text itself, no matter how well-written or accurate the information may be.
Readers will naturally assume that you wrote the text yourself if you leave out the source. I think this is the main issue when someone posts AI-generated content without quotation marks and a source. Simply select the text, click the quotation mark button, and cite your source—no more complicated than the example below. If you skip this step, you are either trying to take credit for something you didn't write or you don't care much about the accuracy of the information.
Written by ChatGPT:
Adding quotes and a source to your AI-generated text without writing any additional text yourself may still be considered low-effort, but at least readers will know the text may contain false information because it was generated by AI.
On another note, using generative-AI to translate or rewrite texts in English doesn't seem to trigger detection tool that much. I wrote the first 2 paragraphs of this post in Dutch and asked ChatGPT to rewrite it in English with correct spelling and grammar—I wonder if you, the reader, had noticed. Most generative-AI detection tools concluded the text was likely written by a human.
All free generative-AI detection tool websites that Google listed on the first 2 pages.
There are plenty of websites that help you detect AI-generated text: Google results. Schoolteachers often use such tools to verify written work of students. AI detection software is not foolproof and gives you a probability percentage, so there's no definitive proof someone used generative-AI. But if a user shows a pattern of posting texts with a high generative-AI detection rate (low effort) then I think it's fair to remove those posts if a moderator also considers the content to be of low quality.
I'm aware of those tools but last I checked (not sarcasm, I mean I haven't checked in a while) "not foolproof" is incredibly generous as they regularly flag non ai things as ai. It will get worse as the ai gets better at sounding human / varying styles and kids write more and more like the ai they're copying.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
