If AI was the answer what was the question, again?

Conference on Tue, Nov 5, 09–18 at AdBK Munich, Akademiestraße 2, Main Building, Central Lecture Hall, free admis­sion

With: Elisa Giardina Papa, Orit Halpern, Francis Hunger, Vladan Joler, & Evelyn Kreutzer, Kevin B. Lee, RYBN, Felix Stalder, Hito Steyerl

The polit­i­cal econ­o­my of AI impacts on both its gen­er­a­tive out­put as well as on art scenes and cre­ative labor. Which strate­gies and tac­tics can artists employ in a  sphere of cor­po­rate gen­er­a­tive media to avoid redun­dan­cy and their own obso­les­cence?

The con­fer­ence tack­les these ques­tions through mul­ti­ple per­spec­tives: Theft and extrac­tion of data from users to train large gen­er­a­tive AI mod­els. The replace­ment of human labor, or the sim­u­la­tion of machine labor through human labor. The envi­ron­men­tal impact of heavy com­pu­ta­tion.

How do these become vis­i­ble with­in gen­er­a­tive out­put? How do they mod­i­fy the con­di­tions of artis­tic labor and what are strate­gies of resis­tance against cor­po­rate extrac­tion? How can art and art edu­ca­tion react in the face of the seri­ous chal­lenges to arts‘ aes­thet­ic and mate­r­i­al envi­ron­ment?

Abstracts

Felix Stalder – Unreal Creations. Generative AI as premonition

It is often said that gen­er­a­tive AI is con­ser­v­a­tive, even reac­tionary, that it can only recre­ate ver­sions of the past, con­tained in its train­ing data. That data, it is also reg­u­lar­ly point­ed out that, con­tains main­ly his­tor­i­cal bias­es which gen­er­a­tive AI repro­duced and ampli­fies. Both of these claims are cor­rect, and trig­ger calls to make AI more fair, account­able and trans­par­ent. In my talk, I want to shift away from such ques­tions of rep­re­sen­ta­tion, and focus on the gen­er­a­tive dimen­sion. Generative AI pro­duces “unre­al data” that is, pre­sen­ta­tion of things that do not exist, but might come into exis­tence based on that data. They are pre­mo­ni­tions of worlds to come. Thus, the ques­tion we need to ask, is per­haps less if the images (videos, texts, and sounds) are cor­rect or fair, but whether the worlds they envi­sion are desir­able.

Elisa Giardina Papa – Surrogate Data and Ungovernable Data

In this talk, Elisa Giardina Papa will out­line the the­o­ret­i­cal and archi­va
research which informs two of her video instal­la­tions, Technology of Care and Cleaning Emotional Data. Presenting images she col­lect­ed while work­ing as a “data clean­er” for var­i­ous AI sys­tems, she will address the ways in which machines are dis­ci­plined and trained to see. Tracing, bound­ing-box­ing, and label­ing are key oper­a­tions used to teach machines to sep­a­rate Data from data, sig­nal from noise, and order­ly things from dis­or­der­ly ones. They are also, Giardina Papa argues, the onto-epis­te­mo­log­i­cal oper­a­tions of mod­ern impe­r­i­al and colo­nial con­quest. Ultimately, this talk will be an invi­ta­tion to reflect on modes of see­ing oth­er­wise which remain rad­i­cal­ly unruly, irre­ducible,
and incom­putable.

Vladan Joler – Calculating Empires: A Genealogy of Technology and Power Since 1500

Calculating Empires is a large-scale research visu­al­iza­tion and phys­i­cal instal­la­tion explor­ing how tech­ni­cal and social struc­tures co-evolved over five cen­turies. It traces tech­no­log­i­cal pat­terns of colo­nial­ism, mil­i­ta­riza­tion, automa­tion, and enclo­sure since 1500 to show how these forces still sub­ju­gate and how they might be unwound. In this guid­ed tour, Vladan Joler will deep dive into some of the shifts in com­mu­ni­ca­tion tech­nolo­gies, infra­struc­tures, and com­pu­ta­tion­al archi­tec­tures, and how they are entwined with the his­to­ries of social con­trol and clas­si­fi­ca­tion.

RYBN – Human Computers, a laborious mediarcheological study

“Human Computer” (RYBN.ORG, 2015–2022) is a medi­arche­o­log­i­cal inves­ti­ga­tion that argues that com­pu­ta­tion is root­ed in Adam Smith’s Division of labor, as exem­pli­fied by the man­u­fac­ture of cal­cu­lus estab­lished in 1793 by Gaspard Riche de Prony (D. Roegel, D. A. Grier). The research revis­its sev­er­al sem­i­nal attempts to sim­u­late intel­li­gence: From the infa­mous Mechanical Turk (W. von Kempelen, 1770) to the con­tem­po­rary human-in-the-loop com­put­ing par­a­digm, pass­ing by Eliza (J. Weizenbaum, 1964) and the Turing test (A. Turing, 1950).
It uses the per­spec­tives of dig­i­tal labor stud­ies, as well as Philip K. Dick’s notion of sim­u­lacra. To reeval­u­ate them by fol­low­ing this com­pos­ite geneal­o­gy, we state that what is called today ‘Artificial Intelligence’ inscribes itself in the long tra­di­tion of the appa­ra­tus of labor met­rics, sur­veil­lance and opti­miza­tion (A. Rabinbach). In the age of Artificial Artificial intel­li­gence, Pseudo-AI and Faux-AI, the ques­tion we ask is: what trick made us loose sight of these labo­ri­ous ori­gins so to ingen­u­ous­ly believe that “Artificial Intelligence” was about intel­li­gence?

Kevin B. Lee & Evelyn Kreutzer – The Generative Afterlife of an Problematic Archive

Generative AI algo­rithms can func­tion as inad­ver­tent archives of polit­i­cal­ly and aes­thet­i­cal­ly prob­lem­at­ic or illic­it images such as extrem­ist pro­pa­gan­da and vio­lent or porno­graph­ic con­tent. Certain search terms can be used to access or in turn cre­ate such archives, yet they are sub­ject to algo­rith­mic trans­for­ma­tions that enforce safe­ty on the one hand and rein­force aes­thet­ic ide­al­iza­tions on the oth­er. This pre­sen­ta­tion shares research relat­ed to this field of ques­tion­ing, focus­ing on two case stud­ies, an upcom­ing film and an upcom­ing video­graph­ic inves­ti­ga­tion, both of which explore how the cur­rent state of online media shape the mem­o­ry, ethics, and rewrit­ing of vio­lent his­to­ries.

Orit Halpern – Speculative Nets: Artificial Intelligence, Finance, and Reactionary Politics

The talk will exam­ine the rela­tion­ship between psy­chol­o­gy, neo-lib­er­al eco­nom­ic thought, and tech­nol­o­gy since the 1970’s. I will dis­cuss how ideas of democ­ra­cy, free­dom, agency, and deci­sion mak­ing were recon­fig­ured in terms of self-orga­niz­ing sys­tems, com­mu­ni­ca­tion, and non-con­scious­ness. I argue this change con­tin­ues to inform con­tem­po­rary pol­i­tics, and shape how we under­stand insti­tu­tions, ‘the human’ and tech­nol­o­gy.

Directions

AdBK Munich, Akademiestraße 2, 80799 Munich, Main Building, Central Lecture Hall

Closest Metro Station: “Universität”, with lines 3 and 6. From main sta­tion use S‑Bahn and change lines at “Marienplatz”.

The venue is wheel chair access­able.

Thank You

Class for Emergent Digital Media, Vasili Vikhliaev, Robert Oeckl and Team, Thomas Köhler, the lec­tur­ers, the AdBK staff, and all vol­un­teers. Funded by the Hightech Agenda Bayern.

Concept: Hito Steyerl, Francis Hunger

Organisation: Francis Hunger, Anja Lekavski

Graphic Design: Katharina Köhler, ktell.de