![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
thenewbuzwuzz posting in ![[community profile]](https://www.dreamwidth.org/img/silk/identity/community.png){kind=small}
su_heraldWe've occasionally linked outputs of things like predictive text engines in the past, especially back when those were an interesting novelty. Now that pretty much the whole world is using ChatGPT, the time of novelty is past, and it's time to clarify where we stand. A summary of this post will be added to the Herald's FAQ for reference.
As a rule, we will not knowingly link AI-assisted stuff unless it's exceptionally original compared to what we've seen before (for example, an entirely new format) or the human artist/author has modified the generated material in some creative and non-trivial way (for example, by using AI to create a background, which they then heavily modify by hand), or the post has sparked substantial discussion.
If someone posts machine-generated material and doesn't label it as such, we--in all honesty--might not be able to tell. However, if we do find out, especially if there's evidence that little to no human editing went into the finished work, we'll remove the work in question from the Herald if we've already linked it, and we'll avoid linking anything else posted by the same person in the future.
We arrived at this approach because the Herald pretty much exists to promote what fans create and what fans have to say. Large text models are not Buffyverse fans, and if their developers have any opinion on the series, it's not apparent in their work.
On a more serious note, one of the Herald's stated purposes is to encourage fandom members to participate in their community. We want to support people pouring time, effort and love into what they create, expressing and discovering themselves in the process, starting conversations based on their own opinions, and collaborating with actual human beings. The process is crucial; it's part of what builds the community. Sure, AI is efficient in the sense that it can provide something vaguely acceptable faster than humans can. But if efficiency was our goal, none of us would be here taking our sweet time to ruminate over a 20+ year old show.
I bet many of our readers don't need us to explain that there are issues with AI, but are more surprised that we're not taking an even more categorical position against AI.
To stay at all relevant and useful, because as a newsletter started in the oughts we are a relic of ages past, the Herald has to make an effort to be open to different forms of fandom participation and keep up to date with new developments. We're well accustomed to linking some things we personally don't like, because our work isn't about each editor's individual opinions. So we're not closing the door to the possibility that these technologies *could* be used for worthwhile creations. In a few more years or maybe even sooner, if the use and regulation of machine learning continues to develop and forms of artistry get more established, we (or future editors of the Herald) might make different decisions about what to include. Or the fad might blow over and AI use in art might become an anecdote of the past, like that era when people put jello in all their cooking.
P.S. We don't always have time to do substantial research as we gather and select links, so if you notice we've linked something that is for sure not human-made, feel free to let us know by sending sunnydaleherald an ask on tumblr or contacting me (thenewbuzwuzz@protonmail.com or a direct message on the webbed site of your choice where you've seen me around). However, we are unlikely to act on such tips without solid proof.
A very happy ritual sacrifice with pie to all who celebrate!
As a rule, we will not knowingly link AI-assisted stuff unless it's exceptionally original compared to what we've seen before (for example, an entirely new format) or the human artist/author has modified the generated material in some creative and non-trivial way (for example, by using AI to create a background, which they then heavily modify by hand), or the post has sparked substantial discussion.
If someone posts machine-generated material and doesn't label it as such, we--in all honesty--might not be able to tell. However, if we do find out, especially if there's evidence that little to no human editing went into the finished work, we'll remove the work in question from the Herald if we've already linked it, and we'll avoid linking anything else posted by the same person in the future.
We arrived at this approach because the Herald pretty much exists to promote what fans create and what fans have to say. Large text models are not Buffyverse fans, and if their developers have any opinion on the series, it's not apparent in their work.
On a more serious note, one of the Herald's stated purposes is to encourage fandom members to participate in their community. We want to support people pouring time, effort and love into what they create, expressing and discovering themselves in the process, starting conversations based on their own opinions, and collaborating with actual human beings. The process is crucial; it's part of what builds the community. Sure, AI is efficient in the sense that it can provide something vaguely acceptable faster than humans can. But if efficiency was our goal, none of us would be here taking our sweet time to ruminate over a 20+ year old show.
I bet many of our readers don't need us to explain that there are issues with AI, but are more surprised that we're not taking an even more categorical position against AI.
To stay at all relevant and useful, because as a newsletter started in the oughts we are a relic of ages past, the Herald has to make an effort to be open to different forms of fandom participation and keep up to date with new developments. We're well accustomed to linking some things we personally don't like, because our work isn't about each editor's individual opinions. So we're not closing the door to the possibility that these technologies *could* be used for worthwhile creations. In a few more years or maybe even sooner, if the use and regulation of machine learning continues to develop and forms of artistry get more established, we (or future editors of the Herald) might make different decisions about what to include. Or the fad might blow over and AI use in art might become an anecdote of the past, like that era when people put jello in all their cooking.
P.S. We don't always have time to do substantial research as we gather and select links, so if you notice we've linked something that is for sure not human-made, feel free to let us know by sending sunnydaleherald an ask on tumblr or contacting me (thenewbuzwuzz@protonmail.com or a direct message on the webbed site of your choice where you've seen me around). However, we are unlikely to act on such tips without solid proof.
A very happy ritual sacrifice with pie to all who celebrate!
no subject
Date: 2023-11-23 10:27 pm (UTC)no subject
Date: 2023-11-23 11:27 pm (UTC)