AI, Fandom, and Artistic Merit
ORIGINALLY WRITTEN: Spring 2024
NOTES: Jason attempts statistics but manages to fumble hard.
The below is part of a small experiment? I did on Tumblr in March 2024 surrounding the early discussions of generative AI usage in fandom. My goal with the research was to determine if there was a difference in the way people perceived generative AI use for visual art versus written art.
I posted a week-long poll on my main Tumblr blog on March 28, 2024, with the following structure (proper structure edited in; I tend to post in all lowercase but wanted it to flow with this page):
I see a lot of posts about these kinds of things in fandom tags, and across different social medias, so I'm curious bc I never see anybody actually really talk about it in discussions about AI-
How do we feel about Character.Ai and similar websites / applications (re: AI chatbots)? Please read options carefully
The options were based on a Likert-scale and formatted in a manner that worked with Tumblr's vertical poll structure. They are as follows:
I'm generally pro-AI art, and I *strongly dislike* these things
I'm generally pro-AI art, and I *slightly dislike* these things
I'm generally pro-AI art, and I *slightly like* these things
I'm generally pro-AI art, and I *strongly like* these things
I'm neutral about AI art, and I *strongly dislike* these things
I'm neutral about AI art, and I *slightly dislike* these things
I'm neutral about AI art, and I *slightly like* these things
I'm neutral about AI art, and I *strongly like* these things
I'm generally anti-AI art, and I *strongly dislike* these things
I'm generally anti-AI art, and I *slightly dislike* these things
I'm generally anti-AI art, and I *slightly like* these things
I'm generally anti-AI art, and I *strongly like* these things
The poll was followed by the following caption:
Feel free to explain why in a reply/in the tags! I opted to use a 4-point likert scale as it allows me to include a neutral option (re: AI art as a whole), but get a non-neutral opinion on this specific subset of AI-generated content.
Reblogs for visibility appreciated!
The poll ran through its duration, and automatically closed 1 week later. On April 4, 2024, I published the results in a separate text post. In this post, I used my educational background to break down the statistics and provide some personal insight into what I observed, and what I could have done better. The post is as follows:
I FORGOT THE POLL ABOUT CHARACTER.AI AND SIMILAR SITES ENDED TODAY!!!
So, to briefly break the discussion down (and nerd out about stats for a second)…
I recently did a poll asking people, particularly those within fandom and / or who actively participate in fandom, how they felt about AI-generated art and how they felt about websites / applications like Character.AI (which, to my knowledge, is the most popular named option).
The results of that poll are as follows.
GENERAL STATS
Poll Duration = 7 days
Total Votes (excluding my own; even as a poll-poster, i cannot see live results unless i click an option) = 135
Total Likes = 6
Total Comments (excluding back and forth discussion; not counting tag replies) = 0
Total Comments (including back and forth discussion; not counting tag replies) = 0
Total Reblogs (excluding my own) = 9
Total Reblogs (including my own) = 19
VOTING DISTRIBUTION (rounded; you can see total votes, but not votes per option, so these values were manually calculated by yours truly)
Pro-AI Art = 0.7% of total voters (approx. 1 person)
Neutral re: AI-art = 8.9% of total voters (approx. 12 people)
Anti-AI Art = 90.5% of total voters (approx. 122 people)
Unsurprisingly, most voters were Anti-AI art.
Strongly Dislikes Character AI Bots = 41.2% of total voters (approx. 56 people)
Slightly Dislikes Character AI Bots = 22.8% of total voters (approx. 31 people)
Slightly Likes Character AI Bots = 22.1% of total voters (approx. 30 people)
Strongly Likes Character AI Bots = 14% of total voters (approx. 19 people)
Surprisingly, most voters Strongly Disliked Character AI Bots.
I expected more people to be fairly neutral or favorable toward them, especially with how much I see them promoting in fandom tags, on TikTok, and so on.
Then again, there is a sample bias as I could not reach a larger audience, which I acknowledged when initially posting this poll; I knew it wouldn’t go anywhere. This site has become very anti-engagement… for some fucking reason.
Strongest Overlap / Poll Winner = Anti-AI Art, Strongly Dislikes Character AI Bots (39% of total voters; approx. 53 people)
And this is an interesting insight!
As I said when I ran the first poll, many discussions about AI art are focused on visual media (think: digital art, recreating speed paints, creating videos, generating images, etc). I wanted to focus the discussion this time on writing, as the art mediums are viewed differently by different people.
… Unfortunately, the discussion (what minimal there was) did circle back to visual art. There was only one person who explicitly mentioned written art, which was the core context of the discussion. There was another who vaguely referenced art, as well, so it can count as part of the discussion.
This is in no way, shape, or form, a callout of the user(s) who made these points! I’m not mad, or trying to say anything bad about those specific users.
I just think it’s… curious that in discussions specifically about writing and fic, visual art is seemingly still weighed as “more worthwhile” and “more important to protect”.
And I still think it's rather interesting the difference in which people, particularly in fandom, tend to view written art versus visual art. I'm sure there's a plethora that can be said about the ease of consumption, and the trends toward consumerism within fandom spaces, but that's not really a discussion I feel equipped to have at the moment. So, I'll leave it at that and circle to a follow-up post I made almost a year later, January 15, 2025, discussing the obvious flaws with my methodology (in case anybody wishes to recreate this, or I attempt to in the future):
I ran the poll during late March, 2024, when character AI posts were HUGE in almost every tag you’d go into. It was the era of AI chats being showcased on tiktok, with many prominent fandom creators beginning new series about publishing bots for their fanbases to roleplay with.
I definitely think I would have gotten more engagement and more feasible results on it if…
I knew what tags to use (I’m still bad at tagging things, though) I ran similar polls on other platforms (where this form of content was more popularized, I.E. Tiktok) I didn’t specify that it was around written art (because let’s be real, it seems like literature/written content is severely undervalued in comparison to visual media)
So maybe at some point I’ll run it again. Do a comparative survey. Who knows. But yeah, generative AI doesn’t belong in fandom (or anywhere, frankly.
And that sentiment is still there. Fuck generative AI, all my homies hate generative AI!