Machine Minds and Neo-Ludditism, part 2

Your blogger returns to Medialab in Madrid to soak up some practical creative thinking around the advent of Artificial Intelligence. The “Synthetic Minds” event was a fun afternoon with lots to chew on. Foremost for me was an artwork from the new analytic movement of “critical AI studies”, a formidable early result in that scholarly front. And a proponent of neo-Ludditism, an engineer against the steam hammer. (Guarantee: No AI used in the writing of this post!)
Trevor Paglen gave the keynote talk I blogged in the last post here. It anchored a program of short-term residencies at Madrid’s Medialab Matadero called “Synthetic Minds”. These developent projects teamed artists to use AI in new works. I attended some of the presentations of fledgling works in progress, along with talks by the groups’ mentors.
The first evening was disappointing, but I shall not dish. The mentor of this group was into a kind of technomysticism, and the projects concerned mermaids and an AI “basilisk”. I’d call this the Netflix development team. Artists gotta eat.
The second evening was more to my taste. It was kicked off by curator Bani Brusadin, who I’d met at the dinner after Trevor Paglen’s talk. In his talk, Brusadin referenced a posted January ‘24 discussion on the state of AI development at the WEF at Davos. Ugh: “We have way too many things for us to think about and work on when we think about what can make the most money,” says Kai-Fu Lee. This ur-capitalist encounter group talks about the “vast automaton” with its “hidden decisions” that is being built. (“Computer says ‘no’.”)
Brusadin compared this build-up to the earlier revolution in logistics – container ships, Amazon distribution centers. He reached back to Victorian-era thinkers, and cited Charles Babbage on the autonomy of machines. My notes were fast and now barely llegible, but I’ll note that searching on one term he mentioned, the fiction genre of AI horror, is a thing. “Hal” is the Mae West of the genre.
To confront this emergent basilisk, let’s say, whose glance turns (has turned) us all into data, Brusadin declared that we need a new conception of a public sphere. This can be built from “what refuses to be computed”, the overall relationship between the synthetic and the factual. This is a struggle to define what is real and what is possible.

Anatomy of an AI system (Joler and Crawford, 2018)
He cited Laura Tripaldi's '22 book Parallel Minds: Discovering the Intelligence of Materials on "state-of-the-art chemistry, synthetic biology, and nanotech". This feminist engineer gives examples of data app results which escape their models. Brusadin also pointed to the extensive series of podcasts generated by the Medialab program on their website.

Xenosocialist World-Building

A lot of the discourse in these presentations was about intelligences – animal, vegetable, mineral – that escape human and human-modelled AI comprehension.
The first project presented was a science fiction future, positing “feral AIs” and “AI spirits” resulting from the “implosion of a massive dataset” in 2054. (Animal minds and spirituality are un-computable.) The voice of “xenosocialism” is a pirate radio station out to free us all. Sounds cool, very Cory Doctorow. The object of it is to “take the burden of emancipation off of our human shoulders”. If only.
Another project team had built compact run-around devices with a rather fuzzy object of discovering “Latent Intimacies”. They strapped one onto a cat who slinked around the Matadero complex.
There was (is?) a bunch of other stuff (see links below), but the mind-blower for me was mentor Vladan Joler’s presentation of the new map “Calculating Empires”, a “critical cartography of technology and power” (2020).
Joler showed tiny detailed images then zoomed in to comprehensive, indeed massive charts of the development of AI on multiple levels. Joler built this with collaborator Kate Crawford.

The Map Is the Territory

There’s nothing like big maps and colorful charts to give the idea that massively complex systems as they are pictured can be comprehensible. I’ve been entranced by this illusory kind of understanding since the periodic table in my high school chem class. The work of Mark Lombardi and Bureau d'études, and the fabulous chartings of the Iconoclasistas are always a thrill to encounter. That’s art; that’s modern iconography.

"Nuestra Señora de la Rebeldía", by Iconoclasistas (2011)

“Calculating Empires” is a masterwork of the genre. One tweet praised it as life-changing: @NataliaStanusch tweeted: "Overwhelming in the best of ways." The show inspired her to start a PhD on AI and media studies.
AI studies departments are springing up quickly in academic institutions. I’ll guess that many of those which are not engineering courses come out of game design curricula. The ones that come from media studies sectors are taking a more hard tack big picture approach to an emergent technology that is, as Dan McQuillan wrote, "the steam hammer of limited imagination, a solution to problems defined in administrative offices and enforced through predictive boundaries".

Calculating Empires infographic, a small detail

Critical AI Studies

Dan McQuillan, a lecturer at Goldsmiths in London and author of Resisting AI: An Anti-fascist Approach to Artificial Intelligence, is one among a new crop of critical AI researchers.
I found him by searching “Is there an actual AI agent of political resistance?”. That is the question the Xenosocialists answer in fiction.
In podcasts, McQuillan talks about how "AI as an essentially political technology", an "amplification of processes already underway". And, as Joler’s super-chart demonstrates, the outcome of long waves of historical developments. In his book McQuillan asks "what would this way of doing things exaggerate?"
He is far from the only scholar developing a new Ludditism as techno-politics. He points out those original mill-spikers were experts in the machines they sought to destroy. Ergo, an operational understanding of AI is necessary to be able to critique it effectively.
The host of the podcast (@machinekillspod) is studying the development of the insurance industry, and compared the mutual aid societies of the Victorian era which co-existed with the rise of the actuary. “There is no ethic of care” in insurance, it’s just about distribution of risk. Ditto AI. Not here for you, but for your boss.
As Vladan Joder and Kate Amstrong’s mega-map makes clear, AI which relies on the scrapings of the internet to understand human culture has picked up all the bad epistemological habits of the past – racism, coloniality, sexism – and brutality of all sorts which dances on our screens in a demoralizing daily round.
The $64K question is, can there be an AI for liberation? Who is building it? The answer is to come in the next xenosocialist pirate radio broadcast. But let’s not wait for 2054.

LINKS

Medialab Matadero, LAB#03 SYNTHETIC MINDS (ES & ENG)
https://www.medialab-matadero.es/en/programs
Project "channels":
https://www.are.na/medialab-matadero/channels

talk: Davos discussion, "The Expanding Universe of Generative Models"
Centre for the Fourth Industrial Revolution
https://www.weforum.org/events/world-economic-forum-annual-meeting-2024/sessions/the-expanding-universe-of-generative-models/

Bani Brusadin
https://www.cccb.org/es/participantes/ficha/bani-brusadin/37757

The Chronicles of Xenosocialist AI
https://www.medialab-matadero.es/en/projects/chronicles-xenosocialist-ai

"Anatomy of an AI System" visual map
The Amazon Echo as an anatomical map of human labor, data and planetary resources
By Kate Crawford and Vladan Joler (2018)
https://anatomyof.ai/

Kate Crawford, Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence, 2022
https://knowingmachines.org/publications/calculating-empire

This announcement of the Milan exhibition includes a closeup of the map, showing the density of detail
https://www.e-flux.com/announcements/573730/calculating-empires/

talk: “Refusing the Everyday Fascism of Artificial Intelligence (ft. Dan McQuillan)” on This Machine Kills; @machinekillspod #186.; with Jathan Sadowski and Edward Ongweso Jr.
https://soundcloud.com/thismachinekillspod/186-refusing-the-everyday-fascism-of-artificial-intelligence-ft-dan-mcquillan
text: This new book is a taste of the new scholarly front:
Tiziano Bonini, Emiliano Treré, Algorithms of Resistance: The Everyday Fight against Platform Power (The MIT Press, 2024). “How global workers, influencers, and activists develop tactics of algorithmic resistance by appropriating and repurposing the same algorithms that control our lives.”… “It is no longer enough to automate information flows about us; the goal now is to automate our behavior.”
is available open access:
https://doi.org/10.7551/mitpress/14329.001.0001

talk: Another fun lecture from the COVER project at Essex U in UK:
James Muldoon, "Artificial Intelligence in the Colonial Matrix of Power"
https://youtu.be/kp_BDpt9beA

This paper theorises how a system of coloniality underpins the structuring logic of artificial intelligence systems....

Machine Vision Doesn’t See You: Trevor Paglen in Madrid

Trevor Paglen in 2018. He's speaking at the World Economic Forum. Applause is... lukewarm.

I caught Trevor Paglen last night here in Madrid. He gave a talk to support his exhibition, “Behold These Glorious Times!” It’s a video, a thundering big screen presentation drawn from the data sets used to train artificial intelligence applications. He gave a highly informative presentation on the historical evolution of artificial intelligence, from its hubristic claustrophobic 1960s academic workshopping to its present world-conquering ambitions.
Julia Kaganskiy is the curator. She worked some years at the New Museum, bringing along programs in the old building just south of the new one. (It was there just before the improvement, spending time sorting out videotapes from the MWF Video collection for the New Museum's XFR Statiom project -- now very old media indeed.)
Kaganskiy’s essay in the show handout cites the German filmmaker Harun Farocki's concept of "operational images" -- images made by machines for machines, images that are designed to perform tasks. (That prescient term, coined in 2000, now names both a research institute and a recent book.)
In his essay "Invisible Images", she cites Paglen’s description of an invisible world of visual culture, in which "machine imaging systems no longer merely represent the world but autonomously shape it". The looking and interpreting of images now is being done mainly by machines.
Artists’ engagement with AI has been principally in the field of images that we can see, i.e., DALL-E 2 and its cousins. Trevor trained as a geographer. His work is conceptually based. In his projects he tries to get behind the apps, not use them. He described his work as a “geographer of technology”.

"Getting started with Image Recognition and Convolutional Neural..."
For me it was a bit of deja vu. Paglen’s tutorial was the kind of thing one had to go through in the ‘90s-early ‘00s to understand WTF was up with the internet. All those white scientist men with their big mainframes seeking to overcome their social isolation… He spoke of ole Woody Bledsoe, who studied phrenology (of course) to develop facial recognition in the 1960s. Woody was working under CIA contract. (Of course.) His idea was to mock up a standard head and determine identities based on deviations from that.
Paglen detailed the story of “Eliza”, the first chatbot, who – wait, which – hooked its users in a way that we who now use similar bots to contact corporations find generally useless and infuriating. In the 1960s, however, the effect “Eliza” had on users was a rather creepy illustration of the Barthean “birth of the reader”.
Paglen clearly charts how the AI programs move from taxonomy to learning library. He discussed the development of the ImageNet database project (2009) which seeks to map all objects for machinic vision. Somewhat blind, as he pointed out, to the difference between the apples of, say, Cézanne and those of the Agrana corporation.
Vital to the procedures of the ongoing Big Scrape, Trevor alluded to the “click workers” who labor to remove the “horrible content” from these image sets. (One article on this topic is titled: 'Without Our Work, Facebook Is Unusable'; who cleans up your “friends” bedroom? Backyard gravesites?)
Facial data sets, he observed, were built on the faces of prisoners, using FBI and other PD databases. Who are we when this is how machines see us?
On the dystopian side, depending on how you look at it, or whether you stand to make money on it or not – Trevor repeated the claim that half of internet content is already AI generated. He called it "goop”. I have seen instances of this, chasing obscure vectors of content, but had no idea it was so widespread. How can you tell? For now, for text, by its descent into gibberish.
In the question period he envisioned a not-far-off moment when the articles you read are written and illustrated for you, each person receiving a different tailored version. Would this make us nostalgic for mass market mediocrity?
Other prospective nasties include AIs which watch you as you drive and modify your insurance premiums in real time.
All that aside – that is, to set aside the imperatives of late capitalism, surveillance capitalism, which rule us, and under which as Matt Christman channeling the late Mark Fisher observed, our desires and aspirations are “embedded in the quicksand of neoliberal subjectivities” – AI and the Matrix-ready machines which run it are an imperative field for artists to jump in and investigate.
Trevor Paglen is the Pablo Picasso of 21st century conceptual art.

LINKS:

Since being busted out of the center of Madrid, Medialab is re-animating under a new program head, Eduardo Castillo-Vinuesa. Trevor's appearance is part of a series, "Synthetic Imaginaries" looking at "non-human agency". "Synthetic Imaginaries", curated by Julia Kaganskiy
https://www.mataderomadrid.org/en/schedule/synthetic-imaginaries

For the wider arc of his work, here's Paglen talking with Eyal Weizman of Forensic Architecture, both great investigative artists --
https://www.youtube.com/watch?v=vvE-QoMGjck

Exhibition handout – Julia Kaganskiy, "Behold These Glorious Times! (¡Contemplad estos tiempos gloriosos!)" PDF Published on Dec 27, 2023
https://issuu.com/mataderomadrid/docs/programa_de_mano_trevor

Article she cites – Trevor Paglen, “Invisible Images (Your Pictures Are Looking at You)”, December 8, 2016
https://thenewinquiry.com/invisible-images-your-pictures-are-looking-at-you/

"Trevor Paglen: Invisible Images of Surveillance"; 12 min.
speaking at the World Economic Forum (wow!), 2018
https://www.youtube.com/watch?v=ijVTdSoZEC4

“Woodrow Bledsoe Originates Automated Facial Recognition”
https://www.historyofinformation.com/detail.php?id=2126

"ELIZA is an early natural language processing computer program"
https://en.wikipedia.org/wiki/ELIZA

ImageNet
The ImageNet project is a large visual database designed for use in visual object recognition software research
https://en.wikipedia.org/wiki/ImageNet

VOA - Voice of America English News, "Content Workers in Africa Sue Facebook, Report Poor Work Conditions", July 11, 2023
https://learningenglish.voanews.com/a/content-workers-in-africa-sue-facebook-report-poor-work-conditions/7174986.html

Maggie Harrison Dupré, “Short Road to Nowhere: Huge Proportion of Internet Is AI-Generated Slime, Researchers Find Striking”, 1/19/2024
https://futurism.com/the-byte/internet-ai-generated-slime

Side note: Before his talk, Trevor recognized me from old ABC No Rio days. He introduced me to the young curators as an “OG”. In his student days, Trevor was involved in a SF self-organized anarchist punk venue, that was like a kind of sister to ABC No Rio. It had a geographic name, which I simply cannot recall. [Update -- Chris Pramas on FB tipped me to the place. It's Gilman Street. https://en.wikipedia.org/wiki/924_Gilman_Street] The punk scene of the ‘80s was national, international, with lots of travelers then as now.
“Alternative Voices: 1980s Punk San Francisco”
Photographs by Jeanne Hansen / Interviews and introduction by Jonah Raskin
Alternative Voices | San Francisco Public Library (sfpl.org)

Jello Biafra in 2018; icon of the San Francisco anarcho-punk music scene