The Image Generation
Arguments over using AI visuals as a writer
I had an interesting comment on my recent article, “Words,” and (while it is not directly related to the article below) it spurred me to finally tap out this piece, one I have been mentally planning for months; using AI images on Substack.
In the aforementioned comment, subscriber JD Solomon made an interesting observation…
I think the subtitle for this one should be "and sex sells" as indicated by your cover photo.
In one aspect he is not wrong, and I give him props for noting it. However, it is also about more than just offering pleasing visuals for the reader. Indeed, there are usually very specific reasons why I generate - and use - the images that I do.
For example, with the cover-image he mentions: In that article, I discuss specific culprits in the ongoing attack on words. Here is a quote from the article, with the corresponding image beside it…
My goal when generating this photo was to represent certain archetypes. After nearly a half-hour of attempts, the above-image came about; I took one look at it… and just knew I had my cover for that article.
As I said - reasons for everything.
Which brings me to the point of this article, but allow me to drop a caveat. Here I will discuss using AI in an artistic manner, be it visual or written; this does not go into trusting information that ‘chatbots’ provide (I have already discussed that in an earlier piece), nor its use in writing scholarly papers. Those set off different alarms, about which (especially with the former) I have very grave concerns.
With that noted, the use of AI imagery is still a contentious issue, especially here on Substack. Via both DM and email, I have personally received well-reasoned and passionate arguments against it, and have considered them thoughtfully with an active-mind. I understand and appreciate their points-of-view, and value the fact they took to time to engage in meaningful dialogue and not verbal assault.
Along with those, there have also been a few ‘hot-takes’ which were a wee-bit more… shall we say, ‘judgmental?’ I have seen several examples of these on Substack Notes, with one going to so far as to state (paraphrased)…
If you can justify using AI for your images, why would I trust you to not use it for your writing?
I must challenge the spirit of this notion. To be fair, I think such sentiments are undoubtedly coming from an integrious intent (insufferable as they can sound), especially since they will oft go on to encourage writers to instead pay artists to use their original works or (if one cannot afford it) utilize ‘free-to-use’ images from places like Unsplash.
However, that one notion? Whether deliberately or not, it causes a reader to pause and reflect on the basis of their entire argument.
If you can justify using AI for your images, why would I trust you to not use it for your writing?
Because I am a writer - I am NOT a visual artist, nor do I claim to be. And as a writer who generates and uses AI imagery, yet puts relentless (some might say ‘obsessive’) work into every single word I write?
Yep, this needs to be challenged.
To me, the entire discussion revolves around genuine integrity, and that is demonstrated through hard work and being candid with your readers. As an example, at the close of each post here on Substack I have a ‘Notes…’ section, where I offer follow-up on what is written in the article; in every single post where I used images generated with AI, I have included this statement (or some variation of it)…
-- Unless otherwise credited, all images were generated by the author, using Grok 2 (on X) or Substack’s AI Image Tool.
I do not take credit for the artistry found in the photos used - I let people know up-front that I used a tool of artificial intelligence to generate the images, and I detail which tools; I do not claim credit for that which I do not create. I will, however, take credit for every word I write and will clearly mark any text which is not my own, through the use of block-quotes and (when available) a link to the original work.
I do not, and have never, used AI for my writing. Well, except for one AI tool - a tool which we ALL use (even the scribe of the Note I paraphrased above, I’d wager); we’ll get to that in a moment.
So why do I use AI imagery? To do what I can to create the best possible product for my subscribers. As Substack is a business based on quality, and that we live in a culture centered around excellence (theoretically, at least, although Ayn Rand’s ‘aristocracy of pull’ seems more prevalent these days), my job is to offer readers the highest-quality experience I can give them.
The writing is always the first priority, of course, but the visual aspects cannot be ignored. In a writer’s Utopia every success would be based entirely on the words through which they sweat, but - as with every Utopia - that is not the reality. The Internet is both a text-based and visually-based environment (some might even include ‘sound’ in the mix, but that’s a different discussion), and you must address both aspects to satisfy your readers.
Nature of the beast.
As to the other options for image-use? The free images available from Unsplash and others (and I have used many, including Pexels and Wiki-Commons) are oft unusable for the subject matter I cover. While visually stunning on occasion, finding an image which encapsulates the overall theme of my articles is damned-near impossible.
(As to Wiki-Commons, you also have copyright issues when you use an image without crediting to creator; I always do so in the caption on the article itself, of course, but since there is no way to credit them when that picture is used as a cover-photo on the home page? Those too become unusable, at least as covers.)
And in regards to paying artists to use their work as cover-images? I cannot speak for you, but I do not have extra green lying around to spend on that which I can get for no cost (or for a nominal fee)… so that’s not happening. As much I would love to be able to do dole out the scratch for such works, there is no way I can afford it.
And with this comes one of the great things about AI image generation - it levels out the playing field.
Before, the established (or wealthier) writers had all the advantages when it came to exposure; they could trade their name-recognition for artist’s favors - or could afford to purchase the rights - to use the best images or most beautiful artworks for their articles. This allowed them to dominate the visual side of the market, as their products were easily recognizable AND aesthetically pleasing.
Obviously us commoners, without connections or coin? We were hamstrung, and playing this particular game with a distinct disadvantage.
Now, however, if a wordsmith is willing to put in the effort they too can create an amazing overall experience for the readers, without being crippled by unusable or unappealing graphics. While name-marketability is still a factor, at least with AI image-generators? Writers have the tools to manifest the entire package on their own.
So yes, I will use whatever tools I have available to facilitate the best experience for my readers (they deserve that much), and - so being I do not claim credit for that which I do not create - I am perfectly comfortable with so doing. Still have to tell a good story - and still have to tell it well - but as long as you stay true when it comes to the integrity of your writing, and are up-front with your readers about when you do use AI?
In my opinion, you should feel comfortable using it as well.
And now, confession time: I do use artificial intelligence when writing… and, as I mentioned earlier, perhaps every writer online does (with the possible exception of Kent Peterson, who literally types his missives out on old typewriters) - even AI’s most-vocal detractors use it. It is basic AI (if we applied the Oparin–Haldane Theory to AI’s evolution, it would be similar to those first organic compounds in the primordial ooze), but it is still AI.
It is called… ‘spell-check.’
That’s right, folks - ‘spell-check’ is indeed artificial intelligence, and instead of hiring a pricey proof-reader? I use it instead, and do so with regularity.
So, now that my scandalous ‘AI-shame’ is out of the bag, I hope you will still respect the effort I put in and not judge me…
… despite this terrible, simply-awful, no-good admission.
Thank you for reading The Stone Age. To support me and my work, become either a free subscriber or a paid Member! Annual paid subscriptions are now 50% off the monthly rate, and include “Written In Stone” a weekly newsletter exclusively for Members only!
Join me, as we build a unique Substack subculture of information, entertainment, and enlightenment.
Thoughts on the AI ‘information’ realm…
Not ready for a Membership, but still want to support my efforts to grow The Stone Age?
Notes…
-- I struggled with whether to directly identify the individual who made the quote about using AI, but I chose against it; I had no desire to attack (nor even appear to attack) them directly, so I only used the (paraphrased) quote as an example of the opinions that anti-AI-image folks share.
-- Unless otherwise credited, all images were generated by the author, using Grok 2 (on X) or Substack’s AI Image Tool… but I already mentioned that above.
I actively seek out writers who produce authentic content. To be honest, I don’t pay close attention to the attached images because I want the words to tell the story. I think because my father was a Hollywood guy and the other half of my family is involved in politics, I am not interested in visual imagery aside of what I have seen in fine art museums or nature and I loathe everything “fake”. I appreciate your writing because I don’t need imagery when your story-telling is so good. I read a lot, to the point I receive gifts such as coffee mugs that state, “Yes, I really DO need this book” thus, I believe I am able to discern authenticity relatively well. It never entered my mind that you used AI in your writing because I detect feeling in your words and I do not believe AI is capable of emotion.
I've said before and will say again, generative AI is a tool best suited for amateurs. Total novices won't be able to tell when the AI has made a mistake and therefore use it badly, and experienced artists, writers and programmers will do the job faster and better without wasting time fixing the mess the AI has made.
If you have any ambitions in any area you think AI might be of help, tread carefully. It will not make you better at what you're trying to do, it'll only stunt your development.
Otherwise, have at it, as long as you present it honestly and don't imagine you're outdoing anybody with real skills.
Aside, personally I don't see much point in interrupting your articles with fake art. I appreciate when writers highlight artists who are doing visually something related to the content of the article (or have illustrated it intentionally). But an AI image is worth zero words and doesn't really add much other than distraction.