Happy holidays my UMD fam. I've just updated UMD forums so that all replies will now inherit the Synthetic tag of the original post if it exists.
When you start a new thread in the AI Wam forum, the Synthetic checkbox comes pre-checked. But for replies it was not, and it would be up to the person replying to check the Synthetic tag on their reply. I had it that way because I felt that if their reply contained no actual synthetic images (or text), it shouldn't be tagged that way. I also wanted to avoid false positives in case someone happened to reply with non-synthetic images for some reason.
But people routinely forgot to check the Synthetic box even when adding images, causing a lot of moderation work. So now, every reply will just inherit the synthetic tag of the original post. This is how we already handle the gender tagging. Some false positives will occur (replies with non-synthetic images, or male pics uploaded to a female thread), but the friction from that will be way less than the proliferation of untagged AI content that you can't filter against. It should also help search because you don't want to see replies to a subject that you don't like.
Also, the X-rating tag will be inherited from the original post. So when you reply, you will no longer be able to set the X-rating or Synthetic tag, just like you could never set the gender.
I've gotten suggestions to make a policy where all Synthetic / AI content had to be contained within this AI Wam group. First, we generally don't move content into these Groups as they are supposed to be run by UMD members and we don't want to force you to post to them instead of the actual forums. But the larger reason is that AI will only become more ubiquitous and we'll be on the wrong side of history in trying to avoid it instead of embrace it. It won't show up just in the forums. So I think a beefy site-wide tagging / filtering system is the way to go for now (we don't exactly do "algorithms" yet, so filters it is).
Today I'll be going through and refreshing all of the replies and pics in the AI Wam group so they align with their main posts. Going forward I hope this makes things easier for all of us.
Oh and don't forget that there is a new "Content Filters" section in your preferences. https://umd.net/preferences/#content_filters That section will be expanded any time I add a new control for you to use. Next up will probably be the "penis or vagina" setting that will go alongside the Gender filter
I understand the desire to embrace AI generated images. I think in the stories section, there should be a ban. I like putting the AI in their own area. I understand where it can help make up where there are people who's talents are there but there are things they cannot get right. I don't like the idea of replacing the human aspect of art with AI. I am also one of those people that refuses to use self checkout. I will always be against replacing humans with AI. I also acknowledge that I am probably on the wrong side of history with this one. Thank you for all the hard work on this.
Messmaster said: I've gotten suggestions to make a policy where all Synthetic / AI content had to be contained within this AI Wam group. First, we generally don't move content into these Groups as they are supposed to be run by UMD members and we don't want to force you to post to them instead of the actual forums. But the larger reason is that AI will only become more ubiquitous and we'll be on the wrong side of history in trying to avoid it instead of embrace it. It won't show up just in the forums. So I think a beefy site-wide tagging / filtering system is the way to go for now (we don't exactly do "algorithms" yet, so filters it is).
Is posting AI images to my profile ok if the pictures are tagged Synthetic? Or should all AI stuff be limited to this group?
dalamar666 said: I think in the stories section, there should be a ban. I like putting the AI in their own area.... I don't like the idea of replacing the human aspect of art with AI.
AI won't be replacing anything, I don't think. Just as many people will keep craving natural content, but AI will add whole new genres of stuff we can enjoy and experiment with. I don't think we should be afraid to see what people do with it. It's just a new category of tagging and filtering, to me. :splosh
Kabe22 said: Is posting AI images to my profile ok if the pictures are tagged Synthetic? Or should all AI stuff be limited to this group?
Yes. I'm using the term "synthetic" as an umbrella term to refer to AI and any non-AI technologies that can create convincingly realistic content.
Messmaster said: AI won't be replacing anything, I don't think. Just as many people will keep craving natural content, but AI will add whole new genres of stuff we can enjoy and experiment with. I don't think we should be afraid to see what people do with it. It's just a new category of tagging and filtering, to me.
AI already is. There is at least one author that has said as long as AI is allowed to post here, they won't be sharing their stories. So there is one person replaced by AI. Telemarketers use AI in their auto dialing. Zoom is using people's voice and mannerisms without their knowledge because no one reads things. Zoom is using that to train AI. One of the things being argued about with the actors strike is the fact that studio heads want to hire them for 8 hours of green screen and then they have their image, voice and mannerisms for as long as they want to use it. Anytime they want to any place. AI can fill in the rest.
There is another aspect of this that is happening that I like. That new Beatles song is great. But at the end of the day humans would have been able to really do this. But when? Would Paul and Ringo still be alive for it? AI did not replace voices or add voices. The AI separated John's voice from the piano so it could be cleaned up and everything. AI is replacing a future human in that aspect. AI is not art. AI is not imaginative. AI does enhance artistic talent that might not exist. But at the end of the day, if you want those kind of talents you can practice and work hard at it instead of taking a short cut. But that is their choice.
Messmaster said: AI won't be replacing anything, I don't think. Just as many people will keep craving natural content, but AI will add whole new genres of stuff we can enjoy and experiment with. I don't think we should be afraid to see what people do with it. It's just a new category of tagging and filtering, to me.
AI already is. There is at least one author that has said as long as AI is allowed to post here, they won't be sharing their stories. So there is one person replaced by AI. Telemarketers use AI in their auto dialing. Zoom is using people's voice and mannerisms without their knowledge because no one reads things. Zoom is using that to train AI. One of the things being argued about with the actors strike is the fact that studio heads want to hire them for 8 hours of green screen and then they have their image, voice and mannerisms for as long as they want to use it. Anytime they want to any place. AI can fill in the rest.
There is another aspect of this that is happening that I like. That new Beatles song is great. But at the end of the day humans would have been able to really do this. But when? Would Paul and Ringo still be alive for it? AI did not replace voices or add voices. The AI separated John's voice from the piano so it could be cleaned up and everything. AI is replacing a future human in that aspect. AI is not art. AI is not imaginative. AI does enhance artistic talent that might not exist. But at the end of the day, if you want those kind of talents you can practice and work hard at it instead of taking a short cut. But that is their choice.
I think you misunderstand how AI works. Despite claims otherwise, AI is not sentient and doesn't do things on its own. It does what it was programmed to do, and follows human instruction.
I have an AI-assisted story on this site. I had to constantly redirect ChatGPT in order to get the rough story, then had to make a ton of edits and revisions to fit my vision. Writing the story with ChatGPT took maybe about an hour or two. Editing it before posting it took 3-4 hours. I didn't just copy and paste the story from ChatGPT. Try writing a story with it sometime. GPT-4 struggles a lot with descriptions, metaphors, and understanding human emotions. It forgets basic details and is constantly changing previously-established details. For as useful a tool as it is, it's just a tool and still requires a lot of human effort to make high-quality written work.
If you've spent much time in the AI WAM group (or played with image generators yourself), you'll find pretty quick that prompt engineering to get results even close to what you want takes a lot of thought and tweaking. You can tell the difference between those posters who've spent a lot of time figuring out how to manipulate AI and those who just post whatever results they get.
Personally, I've spent hundreds of hours using both ChatGPT and other chatbots, as well as image generators, and I'm just scraping the surface of what AI can do. But it can't do much at all without input from real humans, whether it's coding/training the AI or submitting prompts.
AI's inclusion in the SAG-AFTRA strike isn't so much about what studios are currently doing with AI, but what studios COULD do with AI as the technology advances. Some of that is real fears, while some other aspects are just fear about what AI might be able to do some day.
Kabe22 said: I think you misunderstand how AI works. Despite claims otherwise, AI is not sentient and doesn't do things on its own. It does what it was programmed to do, and follows human instruction.
I know that it is not sentient. That it has to be programmed or instructed.
I have an AI-assisted story on this site. I had to constantly redirect ChatGPT in order to get the rough story, then had to make a ton of edits and revisions to fit my vision. Writing the story with ChatGPT took maybe about an hour or two. Editing it before posting it took 3-4 hours. I didn't just copy and paste the story from ChatGPT. Try writing a story with it sometime. GPT-4 struggles a lot with descriptions, metaphors, and understanding human emotions. It forgets basic details and is constantly changing previously-established details. For as useful a tool as it is, it's just a tool and still requires a lot of human effort to make high-quality written work.
If it takes all that effort to fix what is written by ChatGPT, why not just write the story yourself and improve your writing skills? Does ChatGPT really add that much?
AI's inclusion in the SAG-AFTRA strike isn't so much about what studios are currently doing with AI, but what studios COULD do with AI as the technology advances. Some of that is real fears, while some other aspects are just fear about what AI might be able to do some day.
Studio heads made statements saying they wanted to do that with actors. It was not a perceived fear of what could happen in the future. It was also part of the writers strike for the same reasons. What is the point of cultivating creative talents if you can just plug a basic idea into ChatGPT and then spit out a script in a very reduced time scale?
Using ChatGPT is just lazy writing. It is also a slap in the face to people that have real talent that they have cultivated and shared with others. In any situation where an automated entity is seen on the same scale as legitimate writing we are losing as a society. If people want to write AI assisted stories or just straight up AI stories, that is fine. But don't put it in a section with stories written by people with talent and pass it off as being equal. That is just not the case.
What we see in stores is a great metaphor for where we are headed. When the Walmart first opened near me in the 90s they had 20 lanes opened and staffed with cashiers. Then slowly that number was cut in half. Then the idea of self checkout came in where you only needed 1 human for 6 registers. They took out a bunch of lanes to make room for self checkout. Now there is about 15 lanes, of which 3 to 4 are normally staffed with people. Had there not been a need to sell tobacco products there would be less. What happened to those 10 or so jobs? Some of those people had to find work in other professions. Some had to take lower paying jobs because of it. Believe it or not counting back change and watching out for being short changed is a skill that is slowly disappearing.
Messmaster said: AI won't be replacing anything, I don't think..
AI already is. There is at least one author that has said as long as AI is allowed to post here, they won't be sharing their stories. So there is one person replaced by AI.
In that case, AI didn't replace them; They removed themselves out of some sort of fear or resentment. You can't count that against AI itself. Anyone who still chooses to share their human-created work will always be able to do so, and people who like it can still enjoy it. It's not a zero-sum game.
dalamar666 said: Telemarketers use AI in their auto dialing. Zoom is using people's voice and mannerisms without their knowledge...
You're taking the fear of an AI world takeover and applying it here, as if AI content will replace human-created content, and it just doesn't fit. Businesses of course will choose cheap AI over human counterparts all over the world, but that doesn't apply with content creation where it can all exist together. People leaving over a bias against AI doesn't count as AI forcing them out the way it does in the workforce.
Messmaster said: In that case, AI didn't replace them; They removed themselves out of some sort of fear or resentment. You can't count that against AI itself. Anyone who still chooses to share their human-created work will always be able to do so, and people who like it can still enjoy it. It's not a zero-sum game.
You can absolutely count that when the author specifically called out why they were no longer sharing their content here. My point still stands when you take artificially created content and try to claim it as equal to human created content it is a slap in the face to the humans with talent. When you have an item available, then it is no longer available because of another item, the new item replaced the old item.
While I am against AI being used to replace talent, I understand some people like that and should be able to enjoy it. I think any content where AI is used should be required to carry the Synthetic tag.
dalamar666 said: If it takes all that effort to fix what is written by ChatGPT, why not just write the story yourself and improve your writing skills? Does ChatGPT really add that much?
Because I struggle with actually completing my stories. Nobody around me wants to sit down and discuss my ideas and how to turn them into a real story, and professionals expect payment to do that. I don't have the finances to hire an editor or "idea man" to help keep me focused and see the project through. ChatGPT provides that feedback and helps me out when I hit writer's block.
Actually writing the story by myself, including all the distractions due to my ADHD, could take ten times as long to actually finish the story, if I ever got back to it after wandering off. Having somebody or something to respond to and interact with helps keep me focused. I love writing, but I have a very limited attention span without something bringing me back to the story. So, yes, it really does add a lot to my ability to write, even if half of what it writes is nonsense.
dalamar666 said: Studio heads made statements saying they wanted to do that with actors. It was not a perceived fear of what could happen in the future. It was also part of the writers strike for the same reasons. What is the point of cultivating creative talents if you can just plug a basic idea into ChatGPT and then spit out a script in a very reduced time scale?
Once again, that is a fear of future events, not something they are currently doing. "Wanting to do" and actually doing are very different things. SAG-AFTRA wanted to put guidelines in place to prevent this from happening.
dalamar666 said: Using ChatGPT is just lazy writing. It is also a slap in the face to people that have real talent that they have cultivated and shared with others. In any situation where an automated entity is seen on the same scale as legitimate writing we are losing as a society. If people want to write AI assisted stories or just straight up AI stories, that is fine. But don't put it in a section with stories written by people with talent and pass it off as being equal. That is just not the case.
Again, you misunderstand how chatbots work. Read my AI story. Then read my other stories and tell me how much of a difference there is in writing styles. The AI-assisted story is more focused on the events of the story itself with out all the tangents like my others, but the overall quality is the same. Because I rewrote what ChatGPT gave me in my own words, expanded on things the chatbot couldn't understand or get right, fixed all of its weird metaphors and analogies, and added more of a human perspective. If I hadn't said it was AI-assisted, no one would ever have known.
I can understand your argument when it's applied to simply copying and pasting a story straight from the chatbot or app, especially the short ones that were clearly just the response to a single prompt. But calling anyone who uses AI to assist in their writing lazy and untalented? There are hundreds of reasons why people use AI, and if you look at some of the people here who have embraced AI, lazy and untalented is not at all how I'd describe them. Some, maybe. But then you have the ones who have researched and experimented with various AI to do things that you and I could never accomplish. AI is a tool, and you get out what you put in.
dalamar666 said: What we see in stores is a great metaphor for where we are headed. When the Walmart first opened near me in the 90s they had 20 lanes opened and staffed with cashiers. Then slowly that number was cut in half. Then the idea of self checkout came in where you only needed 1 human for 6 registers. They took out a bunch of lanes to make room for self checkout. Now there is about 15 lanes, of which 3 to 4 are normally staffed with people. Had there not been a need to sell tobacco products there would be less. What happened to those 10 or so jobs? Some of those people had to find work in other professions. Some had to take lower paying jobs because of it. Believe it or not counting back change and watching out for being short changed is a skill that is slowly disappearing.
Walmart and other stores that fully embraced self-checkout are starting to second-guess the approach due to increased theft due to mis-scanned or unscanned items, along with other reasons. On top of that, Walmart doesn't exactly offer high wages to cashiers and stockers. As a comparison, McDonald's pays up to $4 more an hour than Walmart where I live.
dalamar666 said:
Messmaster said: In that case, AI didn't replace them; They removed themselves out of some sort of fear or resentment. You can't count that against AI itself. Anyone who still chooses to share their human-created work will always be able to do so, and people who like it can still enjoy it. It's not a zero-sum game.
You can absolutely count that when the author specifically called out why they were no longer sharing their content here. My point still stands when you take artificially created content and try to claim it as equal to human created content it is a slap in the face to the humans with talent. When you have an item available, then it is no longer available because of another item, the new item replaced the old item.
While I am against AI being used to replace talent, I understand some people like that and should be able to enjoy it. I think any content where AI is used should be required to carry the Synthetic tag.
Choosing to remove yourself from the game because you don't like the game doesn't mean the game is bad. It's a form of protest, but that person's choice was just that, a choice. They weren't forced out. They chose to stop posting.
The point of this whole thread was to notify AI users that all AI content now needs to be tagged Synthetic, so that's already covered.
You are ok with it enhancing various issues people can face, but not being used as a copy and paste. I am saying that you cannot tell the difference and that is one of the issues to me. I can understand the staying focused part and struggling with that. I go through that from time to time hence why I have so many projects that get tossed and not finished. I know that there are folks in the story section that write things in chapters. Would that approach help you to be able to finish things? You could have multiple stories going at the same time and may write a chapter a month or something similar. Open a thread to bounce ideas off people with. There are enough people in the community that would be willing to provide feedback. Hell, I would join the thread. My grammar and shit aint the best, but I would be willing to give feedback on story lines. I know that there are times crafting people put projects in timeout and come back to them later. The chapters could work like that. Where when you have a flood of inspiration it all comes out with multiple chapters at a time, then during the droughts it spends time in time out.
Cursive, typewriter, word processor, spell check, grammar check, synonym alternatives, predictive text, Siri / Alexa / Dragon speech to text. Generative AI is the next evolution here, a quantum step rather than an iterative step.
While the concerns on displacing humans here are legitimate, the written word does not divulge the tools used to create it. To simply abstain is not a long-term strategy.
Messmaster said: In that case, AI didn't replace them; They removed themselves out of some sort of fear or resentment. You can't count that against AI itself.
You can absolutely count that when the author specifically called out why they were no longer sharing their content here.
AI didn't replace them. According to you, they deleted their own account because AI content exists here, not because AI literally replaced their content. The self-checkout lanes replacing real cashiers analogy just doesn't apply.
dalamar666 said: My point still stands when you take artificially created content and try to claim it as equal to human created content it is a slap in the face to the humans with talent.
I believe it's a mistake to count other people enjoying something they like as a slap in your face. They are not pre-empting anything that you like to see or offending you in any way with their art. It can coexist!
dalamar666 said: When you have an item available, then it is no longer available because of another item, the new item replaced the old item.
Why would that happen? AI content isn't automatically a threat to real / natural content anymore than wetlook content is a threat to messy. It's just a game of categorization and filtering so everybody gets to see what they like. If somebody chooses not to use the filters available to them, and then leaves the site because they're seeing something they could have filtered, what sense does it make to blame the content itself?
dalamar666 said: I think any content where AI is used should be required to carry the Synthetic tag.
That is the policy, and I'm constantly programming UMD itself to make the filters work better and to help people get the tagging right.
Kabe22 said:
dalamar666 said: If it takes all that effort to fix what is written by ChatGPT, why not just write the story yourself and improve your writing skills? Does ChatGPT really add that much?
Because I struggle with actually completing my stories. Nobody around me wants to sit down and discuss my ideas and how to turn them into a real story, and professionals expect payment to do that.
This is what I'm talking about... AI is the new Photoshop and word processor (like shepushedmein said). It's a tool that people will now use while creating content. We better learn to deal with it instead of trying to resist it.
Thank you for starting this very useful filter function. As much as I love technology and ai, most of the time I am looking for images of real human beings.
One simple suggestion would be to regard posting inside "AI Wam" as "tagging" as well.
Right now I still see lots of ai images (maybe untagged) from this group.
Thank you for starting this very useful filter function. As much as I love technology and ai, most of the time I am looking for images of real human beings.
One simple suggestion would be to regard posting inside "AI Wam" as "tagging" as well.
Right now I still see lots of ai images (maybe untagged) from this group.
The Synthetic checkbox is automatically pre-checked for any thread composed for the AI forum and then all replies inherit that tag. It is not forced to stay checked though if the author doesn't want it. If you do see incorrectly tagged images, please take a sec and flag that?
Just my 2c worth. A while back I was posting old pics from Wetfan or whoever on my profile page - some of them were over 40 years old and the admins took me to task over it as some of the models involved might want to move on from stuff that may be "embarrassing" or in some way compromising even though it's heading for half a century later. I obligingly removed all of the content apart from one which is now my avatar and which, incidentally, nobody seems to have a problem with...
Anyway what's to stop some AI expert from using their archive of similar pictures and altering them even slightly so that the original model or producer can't claim that they're actually "their" pictures?