Dr. Michael Levin
Here are some covers (as of 2023) of issues where our papers appear (see complete list of the references where these appeared and the credits for the artwork inside each issue), with art produced by me, our model species, my lab members and collaborators (co-authors), Jeremy Guay, Anat Zeligowski, and more recently, Midjourney.
I can appreciate how AI-generated art is an interesting medium for blog post covers, but frankly the image on the Interface Focus cover seems like an insult to all the others – it’s hard to appreciate it in any depth, and it’s not obvious that it has any artistic, scientific or philosophical intent. In my opinion, particularly for the cover of a journal, it’s tragic to lose the intent and depth of meaning of the images that this AI-generated one apes, and it is therefore essentially a cargo cult version of what it looks like, with the appearance of subtlety, but in fact with all of it sandblasted off for easier mass-production. I think part of this question is one of aesthetics, and that AI-generated art is capable of very easily looking like a scientific article cover, but I know that I would prefer vastly cruder but more personal images.
Everyone has their preferences, although that image actually does have symbolic content that relates to the themes of the special issue (and it was chosen by the journal out of a bunch of others, most of them not done with AI involvement). More generally, such images arise from a creative prompt by the user, followed by embellishment by a tool (the AI), and then rounds of crafting a better prompt and modifying the final image in various ways. How much and what kind of tools artists should use is an interesting spectrum to discuss. Is Photoshop allowed? Software inside cameras? Speaking of cameras, here’s Picasso dissing photography as a medium:
“In every photographer there was a painter, a true artist, awaiting expression. … Now at least we know everything that painting isn’t. … Two of the most frustrated trades are dentists and photographers – dentists because they want to be doctors, and photographers because they want to be painters.”
There were others like this that I can’t recall off the top of my head; basically when photography came about, many artists thought this kind of “aping” would also be a sandblasted version of true art, which should be created by hand. Everyone can set their purity knob where they want, but I would claim that it’s the effect that an image has on the recipient, not reaction to who/what/how that image was produced, that is the interesting thing. I think the true reaction to whether or not one likes an image should be gathered without prior knowledge of what tools were used to make it. I didn’t do the experiment, but I am pretty sure that in many cases where people said they didn’t like an image because it was “produced by a machine”, they would have loved it if they thought a human did it (and couldn’t actually distinguish the origin of images). Personally, I value content >> origin story, but of course everyone has their own opinion on that issue.
Thank you for responding! I agree with your last point, a big part of how I feel about this kind of thing is a visceral reaction, and not one founded in blind analysis of the content (and not one I have any particularly good reason for). I think the real issue I have, which I could advocate for, is that I would prefer for people to put effort into making the content I consume, and I would prefer to consume content made by people, partly because I believe there is likely more benefit to me from consuming it, but also because I believe there is benefit to the world from it being made in the fist place.
In this regard, I think AI-generated art is more of a challenge than photography/Photoshop/animation software (although I appreciate there are strong similarities in the reactions to each of them as they appeared to AI-generated art), as it poses the problem of allowing a “cargo-culting” amateur (that is to say, someone who wants to get a result which looks compelling to them and doesn’t care so much for the value behind the surface appearance) being more easily persuaded of the value of what they are doing. (it is worth noting I am absolutely an amateur in this area and certainly prone to “cargo-culting”).
There’s an analogy I think here to using pre-made loops in music, where it’s possible to make something convincing without “saying” anything, which is not a situation you are as likely to find yourself in when everything must come from first principles.
I am very interested in this and other forms of ‘lo-fi’ art (lo-fi in the sense of not requiring thousands of hours of technical skill to produce content acceptable for publication, and allowing someone without that skill to nonetheless express their ideas through the medium in question) for blog post/article covers: I believe AI art is an example of this, with the hazards I have mentioned already, but some other compelling options are pixel art, art composed exclusively of a few flat geometric shapes, drawings with thick markers, chiptune music and photography. I wonder if you have explored any of these as alternatives, or any other mediums with similar qualities? (please note I am not saying that e.g. photography or line art are simple and not capable of developing enormous skill in, but rather that an amateurish photograph is likely to be a better candidate for an accompanying image than an amateurish oil painting)
Interesting points indeed! I also wonder about other unconventional sources of art: chimp/dolphin/elephant paintings and music, and similar coming from neural organoid and cells in culture. I am pretty sure I could make a culture of slime molds or human brain cells in a dish that produced output signals that controls a digital or even mechanical arm with a brush and makes things that looks very much like elephant art (which a lot of people seem to love). We could make it a perception-action loop where the resulting art was fed back to the culture as stimuli in real-time, just like a conventional biological artist. Similar, or no?
I like that idea, I think the aspect I would be most compelled by is the idea that the culture “likes” what it has made or is “compelled” by it, which sounds a lot like the perception-action loop. Do you know if there are examples of simple lifeforms making images of things they like, like monkeys making fake bananas or (in this case) a cell culture somehow being incentivised to produce something which gives it desirable sensory inputs? This sounds more like a hedonistic kind of art, if it can be called that, but to me it’s the most obvious way of getting something visual made “intentionally” by a simple lifeform.
Proudly powered by WordPress | Privacy | © Dr. Michael Levin 2026
Leave a Reply