To prevent the Unpopular Opinion thread from being derailed, I’ve decided to make this one for the discussion of Machine generated imagery, or ‘AI Art’, which many characters on Argent Dawn now utilise as visual aids or description authors.
My personal opinion is largely analogous to this, but I’m going to avoid talking about my -political- objections in this thread:
My Opinion on AI Content
Generally put, I think most AI-Generated visualisations lack the personality, individuality and charm of conventional, traditional or digital artwork. Characters, more often than not, resemble soulless dolls to me, rather than pretend-people.
I personally cannot see the value in AI Imagery, with the possible exception of using it as a research tool in early stages of concept art - for example, ‘how might this unusual mixture of aesthetic elements be blended?’
The Ramifications of LLM
The Data-Aggregation Systems we call ‘AI’ might be here, and here to stay. But that doesn’t make it right.
Besides the enormous environmental cost, I believe it erodes at our abilities to independently think, create flawed things for ourselves, and to learn and independently question.
I believe the long-term costs outweigh the short-term gains, whether money is saved or not.
Cylons: Or When LLM Overtake Creativity
A few months ago, I was roleplaying in a small group of people when one character accidentally copy-pasted ‘ChatGPT Says:’ into their dialogue box. It became clear then that rather than embodying the character they were roleplaying, they were utilising the LLM to generate their dialogue and descriptions.
It was shocking for me at least, as not only was it hard to spot, but in essence, it meant I had not been roleplaying with a real human being, just what the algorithm believed I, and the other player, wished to read.
I couldn’t help but be reminded of the Cylons from the 2005 version of Battlestar Galactica. Synths from Fallout 4, if you’re younger.
While I recognise this instance was rare, I nevertheless found it troubling, especially given the amount of effort and care I personally try to put into writing my characters and then playing as them.
However, I can’t help but wonder if years of entrenched elitism, judging the flaws of other roleplayers, valuing those with art over those without visual representation, is perhaps why there is this drive towards a kind’ve sterile, machine-cut perfectionism. People wouldn’t use LLM or Content Aggregation to create for them unless the conditions were that they felt the need to.
A Hypothesised Solution
If we were a kinder, more sensitive community, more tolerant of others’ flaws and idiosyncrecies, more accepting of chaos and more welcoming of the outcast, I don’t think AI-generated material would have half as much appeal as it does.
I don’t personally believe we should sacrifice our imaginations to eldritch data-churners owned by unaccountable oligarchs. We should create, and we should fail, and make mistakes, in our journeys to do so.
Share your opinions below.