Love Maureen Dowd. I subscribe to NYT. And The Guardian. And Wiki. I firmly believe in supporting where you dwell.
I will find this and read. Will be interesting to compare it with the literati differentiation in dynastic China.
So much reading homework!
This point you made earlier got lost in the subsequent flow of exchanges - not uncommon on SF, but it would be interesting if you could post something on how you think this form of AI might be used in a museum or educational context.
Hi Mark, I apologise in advance for the length of this response, perhaps this is not the venue but the topic is so interesting to me.
As someone with a museum background, I find the potential for AI in museums and galleries and educational contexts to be game-changing and enormous. Museums today, and traditional art galleries, are undergoing a bit of an identity crisis. They are seeking to remain relevant and keep visitors interested in their traditional collections. Its no longer enough to show the works with small explanatory labels. People want to be entertained, inspired, challenged. That is why you will find exhibitions now with “topical” themes, “re-imagined” collections, “blockbuster” travelling exhibitions, digital art, a massive marketing and merchandising focus, and a move to the online world. AI is going to play a crucial part in this.
There are two main aspects to this museum/gallery AI; the visitor interface and the background management.
Visitor interface will be the attracting factor, with AI enhancing and expanding the “aura” of the authentic object which the gallery/museum collection is based upon. Via AI robotics, chat bots, digital assistant, chipped cards, gallery surveillance, personalised tours, and measuring visitor response and preferences. The entertainment factor will be huge; eg your face on the Mona Lisa, choose your favorite works for an individualised online gallery space to take home and be projected on your walls, deceased celebrity gallery “tours” combining literature, research and art (think Lord Byron’s tour of The Gothic Genre, Gertrude Stein on the 19th Century Paris Bohemian art scene). What we could do now, but MORE, BETTER and with more DRAMA, and more possibilities of visitor interaction.
This is already happening in the digital realm; the projecting digital art light/sound shows that expand/contextualise works, blend images, or even create new melanges of works projected in a flowing format on walls, buildings and gallery spaces. (Here are a few examples, with Chagall and Van Gogh),
and
Van Gogh at the Atelier des Lumières in Paris.
For background management, galleries and museums are basically information mines, and AI will mine them for new connections, expanding, contextualising and contrasting information. Collections will be linked worldwide. Visitor responses will be AI monitored and gauged as to popularity and preferences and maximising museum profit and returns. And all of the technological manipulation and reproduction, far from reducing the “aura” of the original object as Benjamin predicted, will only increase it and make it MORE valuable. It only takes a look at the obscene prices now paid for some of the most sought-after art to see that. So basically, museums and galleries have a huge interest in expanding their “value” to society.
As far as education, we can hope that AI will revolutionise it. Information will now be universally available and personalisable. For example, radically different student learning styles and abilities and conditions will produce individualised learning plans and curriculums. The tiresome paperwork will (hopefully) be removed from the teachers’s burden, allowing them more time to actually teach and less time to administrate and monitor. The automation of administration will also have a huge effect on the running costs of education as well as of museums and galleries. We can hope.
There are also the possible effects of AI on research in general, which are huge, but that is another subject!
Of course, having said all that, I immediately think of the possible negative impacts of AI to museums and galleries. It could lead to the “over-processing” of art, a “Disneyfication” effect. The last time I went to the NGV of art in Melbourne, they had an atrocious exhibition of a 19th-century style large gallery hung ceiling to floor with gold-framed paintings, but with a periodic light show featuring seagulls, moody music, darkness, lightning and other kitsch sound effects, with the random lighting up of a painting or two, meant to supposedly “re-imagine” the traditional art gallery, which made me quite furious. I ended up in the gallery back room hallways, enjoying squinting at badly-lit classical paintings, with tiny labels. I dont like my art “pre-chewed”. And much as I enjoyed the Lascaux facsimile and all the additional educational intrepretations, it in no way compared to seeing the simple, “real thing” at Peche Merle. C’est la vie!
The fun has already begun!
Although, I would rather recommend a trip to the Dalí museum built to himself at Figueres, a short train ride from Barcelona.
More talk on the possibilities of using AI in museums
I agree. Maybe a Surrealist like Dali is not so odd to “revive” via AI, but the kitsch factor is still there. And I would rather see his own museum too.
I guess, like everything, the AI effects will depend on how it is used, and the program, and the development of the genre. Remember the “Gee Whiz” factor of holograms, for example? “Wow, a 3D skull”. Most holograms never got past the 3D realism factor into using the medium in an artistic way. Then of course, developing computer graphics software programs made the old lasers, chemicals and film-based holographic genre pretty much extinct. So we can probably expect that the first AI museum-work will rely on the “Gee-Whizz” factor, but hopefully will develop and be used tastefully and sensitively eventually rather than Disney-ised. On the other hand, many people prefer their art kitsch, too. Can AI develop taste?
Thanks for your interesting, highly detailed vision of AI in museums,
‘The French’ (yes it’s them again) certainly excel at son et lumière at historic sites, but my worry would be that just as at the end of the last century the advent of DTP made large numbers of untrained people suddenly think they could do their own typography. Well they could, but so much of it was bloody awful, with unsuitable choices of font, particulalrly combining different families, illustrations whose aspect ratio had been changed to ‘make it fit’ etc, etc.
I no longer care what happens in the UK (apart from the exchange rate) but still wouldn’t like to see something similar happening in cash strapped regional museums that lacked an in-house designer. Whilst beyond the museum, throughout the outside world, I anticipate a flood of badly produced, over derivative illustrations done on the cheap.
Regarding the AI acronym, from my very limited experience with Stable Diffusion I think ‘intelligence’ is what it lacks because there isn’t any understanding, just digital pastiche.
Mention of ‘taste’ is likely to prompt another, more contentious debate.
But here’s an interesting starting point
18th Century British Aesthetics (Stanford Encyclopedia of Philosophy)
I wonder if that isn’t more a result of the desperation of museums attempting to make their collections relevant for an increasingly entertainment-centric population. The science museum has always been close to my heart, but last time we were there (possibly 20 years ago in retrospect 
) the childrens section had been removed and replaced with some brightly-coloured nominally interactive but quite dumb exhibits. No working model of an Archimedes screw, no Van-de-Graaf generator, none of the other 1001 fascinating items that had been in the basement.
Having bright colours and ‘simple’ exhibits may make the museums more accessible and attractive to the great unwashed, but also takes away from the key role of passing on knowledge in interest to those who aren’t simply waiting to get back to tiktok as soon as they have their phone back.
Another angry old man - with whom I’m in absolute agreement 
That made me chuckle a little.
Up the geeks! Now where are my horn-rimmed glasses? 
Have just had another play with Stable Diffusion, by typing in the title of a work I made about twenty-five years ago, The Museum of Museums (thanks @AnAussie!) The results were unexpected, because SD couldn’t do anything with the repetition within my parameters - I just got pictures of museums.
Below is the 3D Studio maquette for the work and below it is a photograph of the real thing. The sculpture was a model of the neoclassical building in which it was exhibited and whose architecture and configuration corresponded to that of C19th neoclassical museums, which in turn copied the Roman villa that was constructed around a central atrium…
The largest model was placed over a manhole which allowed the successively diminishing versions in its interior to recede below the level of the gallery floor.
The fluorescent roof colour was selected to make the building look digitally generated 5no moss there! @Corona)
Think AI ‘artmaking’ has a long way to go…
Mark,
If you’re interested and your software can handle ifc or if not I can create an NWD file, I have the whole of Notre Dame de Paris in 3D, I’m sure you could have some fun with that.
Thanks, that’s very generous of you, but at the moment I don’t have a reason, or an idea for what I would do with it. Need time to think. However, seconds later there may be something in the essay below that could make good use of your offer.
It’s in French, but you might find some interesting arguments in this short essay that I wrote for my GRETA class - our tutor loved it and wanted it to be published in one of their mags, but everything suddenly stopped due to Covid, I’m not sure if that happened.
La Flèche de Notre Dame: Restoration ou Rénovation ?
L’incendie de Notre-Dame a unifié les Français, mais le concours d’architecture pour la reconstruire a divisé la nation en camps opposés : les restaurateurs et les rénovateurs. Les premiers se considèrent comme des gardiens de la tradition et veulent que le bâtiment soit tel qu’il a été, la flèche déchue de Viollet-le-Duc ayant été recréé exactement; les professionnels du secteur du patrimoine affirment que ce sera également la solution la plus rapide. Toutefois, les rénovateurs affirment que la flèche était un substitut orné du dix-neuvième siècle à une flèche plus petite et plus simple du treizième siècle qui avait été démoli en mille sept cent quatre-vingt six.
Les arguments des restaurateurs sont à la fois populaires et simples: un monument national emblématique restera inchangé et les travaux pourront être achevés à temps pour l’organisation de les Jeux Olympiques de 2024 à Paris. Peut-être aussi, on verra que les Français ont triomphé des forces de la nature.
Néanmoins, Notre-Dame sape les arguments fondés sur l’authenticité et la tradition, qui n’est pas simplement un édifice de style gothique première, mais un continuum architectural qui a subi une modernisation architecturale majeure aux treizième, quatorzième, dix-septième, dix-huitième et dix-neuvième siècles. En outre, bien que le secteur du patrimoine ait défendu la flèche de Le-Duc, les historiens de l’architecture la décrivent comme un pastiche fondé sur son incompréhension du gothique.
En revanche, une tradition citée par les rénovateurs est la longue histoire des gouvernements français qui ont réussi à mettre en place une architecture controversée, mais novatrice, de classe mondiale, allant de la tour Eiffel au Centre Pompidou, à la pyramide du Louvre et au viaduc de Millau. Les rénovateurs peuvent également soutenir que la contribution du verre est essentielle à l’esthétique architecturale gothique chrétienne et moderne laïque. La technologie de verre avancée actuelle peut offrir des solutions très innovantes qui partagent de nombreuses valeurs visuelles avec celles de l’esthétique médiévale de la luminescence. Il convient également de noter que la construction de le-Duc a utilisé des connaissances technologiques de son époque pour créer une flèche plus haute et plus solide que l’originale.
Enfin, bien que les techniques traditionnelles de rénovation et les techniques modernes de construction modulaire puissent permettre à Notre-Dame achevée à temps pour les Jeux Olympiques, le « verdissement » et l’aménagement paysager de le quartier située entre la Tour Eiffel et le Trocadéro, récemment annoncés par le maire Hidalgo, constituent un projet suffisamment ambitieux et plus approprié pour marquer l’occasion.
Nous concluons sur cette base que la meilleure solution consiste à choisir un ajout contemporain excitant à un bâtiment (mélange de plusieurs siècles des dernières idées architecturales) plutôt qu’une préservation sentimentale et simpliste du pastiche historiquement récent.
Unfortunately they opted for the conservative, populist option and the C19th fleche and all the other C19th accretions are now being restored.
I guess, like everything, the AI effects will depend on how they are used
I’m becoming less and less impressed with this branch of AI. Previously I’ve used computer animation processes like ‘tweening’ and averaging to generate Raphaelesque idealised portraits of his muse and mistress that are in evolutionary psychologists’ terms, ‘more beautiful’ than the originals (we’d need an evening, or a book that might appear in a few years time to explain all that, but below are a couple of egs)
But first my source material, Raphael’s portrait of his mistress, known as La Fornarina (‘the baker’s daughter’) to whom we now suspect he was secretly married (ask and I’ll tell you why).
Secondly, Ingres’ neoclassical painting, Raphaël et la Fornarina
And lastly, Picasso’s 1919 portrait of his first wife, Olga Khokova where he riffs on Ingres and Raphael, but only paints in the most important areas of the painting (albeit with superb traditional academic technique).
OK, so that’s the background and I thought it might be useful to use AI to merge the three, but the results were too banal to even bother reproducing, because the programme was algorithmically translating words into images that consisted of pixels and then fusing them into an image without any understanding of anything!
Previously while applying evolutionary psychologists’ reseach on physical aspects of female attractiveness; I’ve used animation and morphing of pre-modern paintings to generate composites that have a higher level of ‘averageness’ than the originals.
’
However I’ve not been using algorithms to find an average, but choosing images to genrerate an ‘average’ THE EVOLUTIONARY PSYCHOLOGY OF FACIAL BEAUTY (ubc.ca)
Unfortunately there’ d be a lot of work trying to achieve what I’d hoped Stable Diffussion might manage in a few seconds - but here’s the probable process - create a set of animated morphs Raphel to Ingres, Ingres to Picasso and Raphael to Picasso then take the mid-point of each morph and eventually create two sets of mid-point morphs that could then be merged into a final image. It’s such a different process and requires a massive amount of human technical intervention (and an understanding of art history) that I feel art historians (and museum staff?) have as yet little to fear from this technology
May be quicker to paint!
Martial Raysse, 1964
Made in Japan - la grande odalisque
Was never sure what that work was hactually about - hovered between earlier art history, Pop Art and Tretchikov’s kitsch .
But wecome to being enlightened …
Beautiful images! Also, great article, thanks for the link. The fact that most AI programs still dont seem to “get” that symmetrical human faces and hands with five fingers are considered beautiful and normal says it all atm. Perhaps there will one day be AI computer artists who can code sophisticated systems like the morphing ones you are using to instantly make interesting and/or beautiful “art” and extend, combine or enhance existing art images. As you say, atm it still requires a massive amount of human intervention and the understanding of quite complex and abstract human concepts. For instance, there is also the well-known paradoxical human fact that while we are most attracted to the “average”, perfect and symmetrical, we can tend to fall deeply in love with someone who has a flaw or two, physically or character-wise. And that ideals of beauty can change with fashions (for example, the gap-toothed smile that is now seen as so sexy but was once a handicap). Could AI ever know the true complexities of the human soul?
With AI, we are probably still in the teenage Marvel/Anime/Video game mode, and judging from the recent films on offer, imagination today is in short supply. Perhaps the screenplays themselves as well as their CGI images are all AI-generated now, based on past successful films/plots. “Special effects” now basically dictate weak plots with awful dialogue, cliche characterisations and predictable situations. (However, maybe that is what the current generation growing up with all this tech want. Even music today is garbage imo. Grumpy old person raises fist to sky…)
And the more I read about recent AI developments in general the scarier it gets. Rather than the threat to art, I’m actually more more worried about the increasing digitisation of everything else in our lives; banking, communication, all our personal information including health and genetic data banks, and how this will be used to control us and will obviously will be hackable. I’ve been waiting for the Luddite reaction to all of this, maybe with the current push towards a global digital currency we will finally revolt. There certainly needs to be some brakes put on all of this technology, we are entering mad scientist mode now.
Not that I could ever presume to enlighten you, here is an old but still good article by Philippe Dagen for Le Monde, full of links to fun
“qualities far removed from classicism: a sense for burlesque and parody, a penchant for outrage and satire, a bitter, sacrilegious humour”
A true artist character.
The Raysse odalisque is fun to see in the flesh. More of a collage with colour and plastic bits of 20th century decorations on an appropriated print of Ingrés’ original. It mocks Ingres’ realist art principles of depicting the ideal, purging imperfections of reality and individuality from art. There is a little plastic fly above that makes it rather funny.
There is more that can be made of the odalisque image of womankind and slavery from the perspective of the 20th and now 21st century. That the piece keeps talking is a good part of arts conversation with us. The colours and light reflections (in the glass bits) make it attractive and desirable to our modern eyes. And I like the coquettish glance that connects the viewer and winks.
All in all, it is defo worth a gander!
Thanks for taking the trouble to explain, I’ve never seen the original and have always tended to be overly dismissive of the Pop Art end of the Nouveau Réalisme spectrum.
As far as I have seen AI is a clever algorithm devoid of emotion that pools together the real creative works of human beings, and puts them together in new ways. I have heard of people seeing watermarks within AI generated art, from where it’s been sampling existing artworks in order to create something new.
On social media I see a lot of people labeling themselves an AI artist, but I don’t see where the creative process lies? You can’t bypass the 10,000 hour rule to really start to master something, and lay claim to being an artist when an algorithm is doing the heavy lifting, and more importantly it sucks all the joy out of the process of creating.
Thankfully as it stands AI isn’t putting down creative traditional brush strokes on canvases (probably just a matter of time though). Within the digital art realm, I can see why artists who have spent years are concerned. It may not impede on their sense of joy in creating, but it will very likely take away their ability to earn… I am concerned that billions of AI generated works will swamp the internet to the detriment of human created artwork, and steal away the artist’s chance to voice their works and find a market for themselves. Sure AI could be a useful tool, but ultimately it’s the creation of billion dollar tech companies, exploiting the works of genuine creation, and will be used and exploited for profit to the detriment of those it took from.
