Generative AI and Ancient History Magazine
So, this is a bit more of a serious topic for the blog today — generative AI.
To be totally frank, I avoid using generative AI in my personal life as much as possible. If I have used generative AI, it has been because I did not know I was using it.
However, given my position as Editor of Ancient History, I need to not let my own personal opinion cloud my judgement when it comes to considering the use of generative AI in the magazine. After all, it could be incredibly helpful. So, I thought about it. I ignored the environmental, social, and political implications of using generative AI — the issues which drive my own Luddite attitude towards it — and assessed it purely on its impact on the magazine.
It is safe to say that we will not be using generative AI in the magazines at all, whether internally or from our contributors.
Firstly, the illustrations. One thing that we at Karwansaray Publishers are proud of is the artwork we include in each issue of the magazine. We believe that well-made illustrations can really bring the ancient world to life — just look at how influential the works of painters like Lawrence Alma-Tadema continue to be. Simply put, AI-generated art lacks the warmth and texture of art made by real people, whether traditionally or digitally. Moreover, we are incredibly proud of the artists who make the illustrations for us. I believe that, in using generative AI, we are stealing from hard-working, independent artists — not only their commissions, but also the very fact that generative AI scans the work of human artists to generate its own images. So, in terms of standards, we will not be using AI-generated artworks in Ancient History.
Secondly, the articles. As editor of Ancient History, I have made every effort to reach out to academic experts in each of the topics we cover in the magazine to ensure that we are giving you the most up-to-date theories and interpretations (and, similarly, non-academic contributors need to show the resources they are working from). Using generative AI to create articles undermines this core part of my editorial philosophy. It has been repeatedly shown that generative AI frequently hallucinates when asked about history, whether in the text it provides or the sources it cites. I am not an expert in every single topic we include in the magazine — no one person can be. Part of the reason I approach experts is that I can trust them to work from up-to-date, reliable material. If we were to use generative AI, I would need to fact check everything it produced for us, which would take up too much time, otherwise the magazine would soon be full of historical inaccuracies and even outright falsehoods. What's more, the 'voice' of AI-generated writing is invariably monotonous. Were we to use AI-generated articles in the magazine, we would lose the diversity of style that human authors bring to their articles, instead adopting monotonous AI-speak, and the magazine's quality would suffer for it.
From a purely production perspective, then, the use of generative AI would decrease the quality of the magazines. Couple this with the environmental, social, and political implications, and I think there is no argument for the use of generative AI.
3 comments
To be honest I feel the main problem with AI for art and perhaps writing is the wild net feed sampling it does, if it had a locked pool of paid references artworks for sampling (I.E. artists who are being credited and paid by being sampled) it would probably get better results with far less controversy.
I am not really sure why writing bots are sampling the internet instead of just peer reviewed scholar sources and as a unpublished fiction writer who isn’t using AI, I am also not sure why you would bother using AI to write outside of instruction pamphlets and maybe very exotic projects, or perhaps as a writing prompter maybe. You might get a few initial book buyers, but bots don’t seem to produce quality fiction writing and people notice, likewise I would trust them even less than human writers when it comes to more academic more because the bot has no adult human judgement and seems very naive. When several AI/SI start blaming the world’s woes solely on Jews, its clear they can’t differentiate the Babylon Bee and Neo-Nazi-Nonsense writing from Academic papers and real life.
Likewise forget “tyrants” rewriting history with AI as rewriting history as already happened more than once in at least the last 250 years in our “Free Society”. Instead imagine mistaking human written falsified propaganda as truth simply because its not AI slop (we already have this problem with some eras as is, for example I would say we take the words of the Ancient Greeks on the whole a little to close heart considering every other civilization that mentions them generally depicts the Greeks as being conmen and liars).
I feel the real places AI would blossom are not really being explored for very questionable reasons.
You are to be commended on your stance regarding the use of AI within your publications. As someone who used to work within the creative industry, being a retired graphic designer, I know of people being effected directly, and very badly by AI.
Thank you! Among the many extreme dangers of AI is the likelihood that tyrants and propagandists will use it to selectively erase and rewrite history. For that reason alone no publisher with integrity should tolerate the use of generative AI.