":')))))))) you realise that gen AI is available to everyone though right??? Queer creators can use it just as much as anyone else??? I just don't understand this post... It really feels like a cheap way to get on the 'AI Bad's bandwagon, and coming from such a thoughtful and insightful creator that's incredibly disappointing... It's okay to not comment on subjects you're not an expert in y'know...?"
Y'all know the drill, I am replying to this publicly but that is not an invitation to send any negative messages to the person I am replying to.
Anyways, let me start by saying that the original context of the post you're replying to is discussing an event where a queer org used generative AI to steal an interview with Keri Hulme. So let's start there. To be clear I don't even know if the original interviewer was queer so let's put the identities of stealer and stolen from to the side. I want to explain the harm done in this example specifically and I hope this is illustrative of what harm generative AI can (and does) do.
The original place I saw generative AI was a queer org that explicitly says they are using generative AI "for good", and as a way to bring more queer history to light. So let's take them at their word, and assume they are not out to cause harm. This is the best example of generative AI that I can imagine, so I hope that makes it clear that I am not coming at this issue from bad faith in any way.
Here is the harm they are causing:
Decontextualizing and rephrasing an interview: I am not going to pretend that I am an expert in academic best practices, but I do believe one thing, if a person is speaking on their own identity and lived experience, it is always much better to directly quote than it is to rephrase. As I read this source, I initially didn't know that it was AI, and I was already upset. An interview that is widely available on the internet with no pay wall, was poorly sourced and made more vague than it was in the initial text. By creating one degree of seperation between the original words of A WRITER (whose literal job was largely based in choosing the right words to describe experiences they had) harm is already done. It makes vague what was once clear, and removes Keri Hulme's voice from her own narrative.
The original interviewer is not paid, or given proper recognition: I get it, sometimes just copy pasting an interview doesn't feel transformative enough, but something that one would learn if they worked in the queer history field and weren't a literal robot rehashing what has already been said, is that not everything needs to be transformed. In those cases, we give credit to the person who said the original words (in this case Keri Hulme), and the interviewer who facillitated the conversation (in this case Shelley Bridgeman). This case (again a best case scenario), takes the attention and byline away from the original interviewer and gives it to an AI.
The original publisher of this story is deinsentivised from paying interviewers in the future: The original publisher of this interview has ads on their website. As a person who also has ads on their website, taking an article like this and rephrasing it for no good reason (the orginal word count was not prohibitive and the rephrasing did not make it more readable), takes money from the publisher. It's pennies, but it's also removing numbers could have been used to justify further interviews with asexual people and archiving of asexual stories. The org that stole from this publication does not interview people themselves so the money and numbers that could have gone to continue to preserve asexual stories goes to stealing them instead.
These are just the active harms that I saw in this specific case. As you said, I am not an expert in generative AI, and will not be speaking as if I am. But I will say that asking me not to speak out on active harm that is being caused in queer history spaces, is disrespectful to my many years in this field.
To illustrate this even clearer: if you were a patron, you would know I recently took down an old article. I have been rereading and editing our backlist of articles, and I found one that no longer fit my standards of sourcing. My standards had recently raised due to a video made by HBomberguy about someone in the queer history space who was stealing from other creators. I watched this video not as a work project, but because I watch most of HBomberguys videos, and this one made me think more critically about sourcing. An AI can't do that. All an AI has is what has been inputted, and it is right now impossible to input every available peice of information about ethics into an AI and get a coherent ethical basis on which it will function.
It is a distinctly human trait to absorb information and change in that way. AI can rephrase information that already exists, steal it, recontextualize it even, but it cannot create something altogether new.
Do I believe that there one day might be an ethical use for Generative AI? Maybe. Do I believe that coming into a queer history space, stealing the words of a Maori asexual author, rephrasing them, and giving the original interviewer and publication no form of compensation for their work, is accomplishing that? No.
On a more personal note: I am coming at this issue with a bias. As a queer history creator, I do not want AI in my space, because it is literally damaging to my financial prospects. It has been like pulling teeth to try and get patrons in the current state of the global economy. I don't blame anyone from that, but I feel very disrespected that I am being asked to compete with a machine now. Not only that, but I am being asked to shut up and be fine with it? No, absolutely not. I cannot and will not stay quiet as space that I have fought tooth and nail to create in mainstream discussions is taken and given to AI.
AI was not supporting me when I was sent gore to try and scare me off of discussing queer history. A person did that. AI was not there to tell me I had written too many sad stories, and I needed some happy endings to remind myself of the good in the world. A person did that. AI was not there when I was being harrassed for supporting and including asexual stories on my website. A person did that.
And after all that, I am being asked to lie down and take it when my ability to pay the people who supported me in those ways, is being threatened. Nope. Not going to happen.
An AI doesn't have to make rent. An AI doesn't understand what it feels like to have to stop holding their wife's hand in public. An AI didn't get calls from people needing comfort in reaction to the election. Pay me for my work, and get this AI nonsense out of my face.
