Countering the ‘AI Ethics’ discourse that has taken root over the last decade—the idea that if we could only get Big Tech’s engineers and entrepreneurs to be better citizens, and design systems with responsibility in mind—the school pushes back against both capitalism and corporatism. And most crucially, against a general lack of imagination about what our lives with AI might be. It also acknowledges that our tableau of algorithmic timelines, natural language processing bots, and bustling data marketplaces is the world we now live in, and pragmatically aims to repair our psyches and social practices—now that AI and automation are so thoroughly entwined with the fabric of everyday life.

As Ganesh and Khan put in their statement: “We want to create a space to consider AI as a prompt to abandon positive and positivist visions of the future, for necessary disregard, for meaningful resistance in a time of an overdetermined computational life. By consciously moving beyond a re-statement of the status quo of solutions, of solutionist discourse, we insist on undoing our own positions.”

‘Undoing our own positions’ resonates, because it posits personal agency instead of (only) pointing a finger at Amazon, Google, or Palantir. The school’s program of panels and workshops is a kaleidoscope of responsibility and resistance, play and provocation, with threads of dreaming, feminist organizing, multisensory stimulation, and whimsical performance running through it. Beyond the savvy sociopolitical framing you’d expect from the assembled brain trust, the school looks like an awful lot of fun.

As the embedded editorial partner, HOLO will respond to the Autumn School’s radical pedagogy and playfulness in kind. Over the coming weeks, this dossier—the AI Anarchies Tracker—will become a rich repository of notes and highlights from its discussions and workshops, but it will also be a place where HOLO pokes, prods, and questions the provocations that bubble through. Beyond our excitement about joining Ganesh, Khan (who we thoroughly enjoyed working with on HOLO 3), the participants, and the JUNGE AKADEMIE team of Clara Herrmann and Nataša Vukajlović, we’re thrilled to share our (un)learnings with all of you.

AI Anarchies Autumn School (Oct 13–20)

The JUNGE AKADEMIE’s AI Anarchies Autumn School, curated by Maya Indira Ganesh & Nora N. Khan. The school is part of the AI Anarchies project, initiated and curated by Clara Herrmann & coordinated by Natasa Vukajlovic. It includes an artist residency programme for six artists in cooperation with ZK/U Berlin and a concluding exhibition in June, 2023 at Akademie. AI Anarchies is supported by the German Federal Commissioner for Culture and the Media.

We tend to think about algorithms and AI as something ‘top down,’ that is imposed on us by a corporation or platform. What are some examples of ‘bottom up’ problem solving or organization that illustrate how people can (or already do) have more agency?

Maya: There’s a very easy way to stop being shown ads about bras, weight loss, light therapy, hot yoga, or whatever else annoys you on Instagram: you can go into the ad settings and (de)select what you want to see less of. Algorithms can feel like magic, and the tech we have can do pretty wonderful things, this is why we love the digital. But algorithms are not necessarily magical. People can and do have agency within the digital by learning how it works, and spending a bit of time in shaping the digital space they want to inhabit.

Nora: I tend to visualize algorithms as ubiquitous: as a cloud, a set of ghosts around us, in the air, the ether, in our minds. Driven, catalyzed, conjured, planted, built by others, much like the companies, institutions, and structures they support and strengthen. Like we learn to navigate and then be slightly anarchic within the physical structures, we do the same with ‘AI.’ People and their machines. We practice disobedience within the nodes, while learning the forest of rhetoric (an old teaching tool), while learning two or three languages. Algorithms are pattern languages with sub-leaves and trees and nodes to cross, with their own practical beauty. Their limits make the world. We practice if-then-maybes at every juncture in our relationships to each other, to these monoliths of institutions. I see agency in this miasma is first found through rhetorical training: learning to map the logic trees. Somewhere in the folds of the map, we find one another.

You both have a wealth of experience in cultural production and education. What have you taken from other schools and institutions, to incorporate into the Autumn School? What have you been careful to avoid?

Nora: Every experience I have had as a student, from my first school (a graduate research lab in a psychology department—the Child Study Center—with double-mirrored observation walls; I have a picture of my 4th birthday party with the mirrors glinting behind me), into middle school, high school, college, graduate school, has given lessons on the myths of meritocracy, mastery, and knowing it all. I’ve learned from many brilliant and generous teachers, but their most imprinted lessons tended to be outside of the classroom: when they’d pull you aside after a critique, or say they could see what you were trying to do.

Those lessons were ephemeral, anecdotal, improper, and always outside the curriculum. They were moments the teacher was vulnerable, a peer, unafraid to show they were learning, making mistakes, and trying new methods along with their students. They were loose. A tender demeanour. I tear up thinking of those moments, the shield dropping. It takes such grace to be intellectually vulnerable. Being loose didn’t threaten their authority; teaching was not about authority! They met you to preserve the thread of an idea, half-formed. As a teacher, working with artists and artists writing, I want to preserve these threads and germs of ideas. I want to stay sharp and recognize them, how they give way to a shift, a turn. AI studies and discourse has long had a violent emphasis on legible mastery. But the critical humanities and art have a pretty strong practice of resistance where needed 🙂

Maya: In a life before academia, I used to work with feminist and human rights-defender movements. The ‘workshop’ has been a key technique in bringing together people working with technology and politics. There was rarely any top-down didactic ‘teaching,’ because everyone was considered an expert in their own right and with something to share. When you make politics, theory, art, and culture, you need particular kinds of spaces for togetherness, and for solitude. You need spaces to play, too. Leadership and direction are essential in taking responsibility for architecting those spaces. Being in nature and by bodies of water is always appealing. Workshops (i.e people!) taught me to think about hosting such spaces in terms of flows, and small experiences of feeling welcome and comfortable; and to foresee the (necessary, positive) frictions that thoughtful people coming together generate.

The school schedule reads like a bucket list of timely provocations and delightfully eccentric workshops. What are some of the more unconventional workshops, activities, or performances each of you are really excited about?

Maya: I’m excited about the many workshops and performances related to sound and listening, and particularly in a moment when visual culture and language culture are being ‘discovered’ by AI industries creating ‘transformer’ technologies. I am also looking forward to Lauren Lee McCarthy’s performance Surrogate. Lee McCarthy’s work confronts the tragicomic, and the tender aspects of being human and inhabiting vast surveillance apparatuses. I believe this work speaks to many of the concerns and questions we have about AI technologies, not just in terms of their extractive or surveillant infrastructures, but also in terms of the absurd conditions of the digital we inhabit.

Nora: We worked on HOLO 3 together, and a review of it (a compendium of essays and works about ‘incomputable’ ideas in relation to AI and machine learning) hit it: many of the pieces don’t seem to be about technology or AI. Here, I’m delighted most by the full, varied range of performers, workshops, and playful strategic sessions, which don’t necessarily seem to relate 1:1 with AI. Maybe this is part of the anarchic piece; perhaps the argument here is that all the things that don’t look like AI do in fact relate to its making, because AIs are myths, imaginaries, narratives, as much as they are algorithms, pattern recognition, data, logics, systems, platforms, decisions. AI—and thinking hard about it and being really confused and frustrated and weirded out by It—creates a special psychological and emotional “situation” in which we’re thinking about what we’re making and what that says about our desires, needs, and perceptions of what constitutes the human, the inhuman, the non-human … I see these workshops as a hydra that collectively try to probe this field. Study is messy and anarchic by nature. My hope is that the workshops and mornings and discussions bubbling in the space … will create emergence—ideas and thoughts we can’t predict.

You’ve been quite careful with language and describe the participants as a ‘study group’ not ‘students.’ How do you see the cohort engaging and contributing to the program, and what kind of outcomes would you love to see from them?

Maya: I think of AI Anarchies as something like a caravanserai, the traditional inns and resting places along the historic, medieval Silk Road that linked the geographies that we now refer to as Europe and Asia. So I think of the study group as fellow travellers meeting by chance, and by design, and having fascinating stories, ambitions, and insights to share. There is so much that AI is reshaping in our human social worlds that we struggle to comprehend and articulate. I think I want to hear new frames and language through which to make sense of things going on. Caravanserais have been sites of cultural cross-pollination and dispersal of ideas-spores, so I hope the study group makes new and stimulating connections that are much more than just *networks.*

Nora: I think Maya has said it perfectly. I’m happy to see the idea-spores dance. Siegfried Zielinski, who was on the advisory panel for the program, early on urged that we keep the anarchic in view, and the meaning of Anarchy in clear sight as we built out the program. And made it feel a bit less anarchic than we had desired. The pressure to formalize the Thing, and our own desires to conclude, make tidy: we had to keep undoing these as curators who like conclusive forms, maybe 😉

You could say designing the school required that we put some of its tentative propositions into practice from the start. We can’t create a space of intellectual play if we’re not practicing that play all along the way. The amazing Akademie der Künste team made this possible, arranging countless calls with speakers and thinkers who’ve already been so generous, and shaped our understanding of what the space could feel like. It felt like a constant rhizomatic negotiation. I see this negotiation and thinking together only continuing, in which we will all work to keep the space open. I think about how spaces for inquiry can close up. Getting amazing people in the rooms, which we feel very proud of doing, is the first step. Maintaining the space, tending to it, responding to the group’s critiques and redirects and evolving needs will be the most profound—and rewarding—challenge. Check in with us along the way 🙂

You close your curatorial statement with the provocation that we should think bigger than ‘what should an ethical AI be?’ and instead dream up “anarchic, strange, and improper AIs.” What are some examples of those kinds of AIs that you would love to see?

Nora: One that plans its own end; one that uses the unimaginably vast bank of personal data gathered from us over a decade absolutely none of the time; as I’ve said with others, forever, a leftist or socialist or anti-fascist AI, however naive that seems. One that takes history and its own history and decisions into account. An AI that works to respond to unseen ‘bias,’ names it, works with knowledge of it, rather than encoding it unreflectingly and more deeply. An AI that’s a genuine confidante who never tells anyone anything. An AI you can barely find and has no interest in what you do and where you consume.

Maya: AI is a set of computational methods and techniques; what might be anarchic and strange and improper are how other methods for being in and knowing human, inhuman, nonhuman worlds might exist alongside and inside it. It is not about better or more appropriate applications of AI, though there are certainly opportunities for this, no doubt. The digital is a source of unending delight and creativity, and many of us are, at this point, part-digital. So I struggle with valuing a place outside the digital as better. I find myself ambivalent and agnostic about this condition, uncertain about the gains and losses.The anarchic-strange is not a point of arrival, but a search for practices of memory, body, collectivity, and fierceness, other logics that can also sustain our hybrid selves. This event, AI Anarchies, is that kind of search; it needs its theories, its culture, its stories; and we are here with this community to do just that.

For your daily dose of absurd, alt- and broken future connoisseurs N O R M A L S will sprinkle the dossier with insights from their AI Anarchy Flattener: a little crawler that scans recordings and notes for all sorts of hidden information to help Autumn School participants know themselves better. In order to perform optimally, this energy monster needs clean, easily computable data: well-formatted speech, reliable behaviours, and as little noise as possible. But worry not: to help participants make their activities more machine-friendly, it will provide simple templates to reduce their anthropic entropy. The bot provokes big questions: Can we, through this cooperation, learn to come and grow together as one? And do we want to forfeit our inner chaos for the sake of better computation in times of energy crisis? Tell it by shaping your arms into a perfect square.

The AI Anarchy Flattener scans daily dossier distillations for hidden information to help AI Anarchies Autumn School participants become their best (most machine-readable) selves. Read more about the N O R M A L S provocation here.

The AI Anarchy Flattener scans daily dossier distillations for hidden information to help AI Anarchies Autumn School participants become their best (most machine-readable) selves. Read more about the N O R M A L S provocation here.

With a double background in computer science and performing arts, Diana Serbanescu works on interdisciplinary approaches to culture, society and technology, with a strong focus on human factors. As team-lead of the group Criticality of Artificial Intelligence at the Weizenbaum Institute, she promoted culturally-aware research on the topics of bias, symbolic power and explainability in machine learning algorithms.

The AI Anarchy Flattener scans daily dossier distillations for hidden information to help AI Anarchies Autumn School participants become their best (most machine-readable) selves. Read more about the N O R M A L S provocation here.

The AI Anarchy Flattener scans daily dossier distillations for hidden information to help AI Anarchies Autumn School participants become their best (most machine-readable) selves. Read more about the N O R M A L S provocation here.