There’s an affective affinity between my internet criticism efforts and the
work of German media theorist Anna-Verena Nosthoff, her partner Felix
Maschewski, and their Berlin Critical Data Lab. After years in the making,
Anna-Verena’s PhD (written in German) was published in early 2026 by Suhrkamp
Verlag in their infamous stw series. Its release prompted the following email
interview, in English, to introduce this majestic 660-page study that covers
eight decades of theory on the “art of digital governance”. Her PhD degree at
the University of Freiburg coincided with the birth of their son Bruno, and her
appointment as junior professor at the Carl von Ossietzky University in
Oldenburg. Time to celebrate, discuss ideas, and steer next steps. Will there
be a long-awaited renaissance of cybernetics, provoked by Europe’s ‘digital
sovereignty’ wake-up call? Is there a need for a meta-theory that brings
together digital media, networks, platforms, and AI? And how many of these
ambitions are holding up in the face of techno-fascism?
Feedback does not equal criticism. Time is necessary to process thoughts,
reflect, and question the existing, then verbalize them in our minds before
standing up to voice criticism. Critique starts with feedback, the central
category of cybernetics. Yet, it is not the case that the other way around. The
philosophical response to the rise of cybernetics as a general theory of
computing in the late 1940s took some years but unfolded steadily in the
following decade—the topic of the first part of Nosthoff’s thesis. After the
initial speculative phase ended, cybernetics as a distinct discipline vanished
precisely as the democratization of personal computing was taking off. Instead
of establishing itself, cybernetics staged its own disappearance, to say it
with Baudrillard. No meta-theory has since taken its place. What replaced the
initial thinking of technical fundamentals was a desperate attempt to
administer its impact. Techno-solutionism is still dominant. While the
reductionism of operational thinking has been unmasked here and there, no
overall alternatives are in place. Nosthoff’s German attempt (including mine)
to reintroduce the ‘criticism’ category should be read as an attempt to uphold
the lack of reflection and regain lost territory.
The aim of Nosthoff is a “genealogical reconstruction of cybernetics
criticism.” The study is divided into two periods: the first deals with the
origins and first- and second-order cybernetics, always with a view to the role
of criticism. In part two, we move from the mid-1980s toward Silicon Valley,
the explosive growth of the Internet, and the dominance of digital technologies
in the present day, when the research no longer revolves around reflections on
‘cybernetics’, but shifts towards a critique of cybernetization (even though
this term is rarely used). This distinction is important because in the 1970s,
cybernetics as a separate discipline slowly disappeared, precisely at the
moment when its impact was truly beginning to materialize and unfold.
Nosthoff describes the historical legacy of cybernetics as a non-disciplinary
meta-approach. Neither a method nor a school, cybernetics refused to establish
itself as a universal science. In the end, it was unable (or unwilling?) to
assert itself. During the post-war decades, criticism still had an object to
relate to in the form of books, journals, conferences, and meetings, when
cybernetics was still reasonably well embedded in the academic milieu; this was
no longer the case later on. From the moment computers became ‘personal,’ the
only thing critics could deal with was the impact, a force so strong and
omnipresent that it disguised its original principles (and the earlier debates
about its premises). In the phase after the mid-1970s, cybernetics criticism
lost its object and was forced to reorient itself. It is remarkable that there
is no clear transition phase—with the exception of the Stafford Beer/Chilean
Cybersyn episode in the early 1970s, so brilliantly brought back to life in
Eugene Morozov’s podcast series.
Since the sweet revenge. How can today’s digital regime be fundamentally
criticized when its founding discipline disappeared fifty years ago? Should its
withdrawal be read as a genius act or even a conspiracy? Regardless, the
institutional poverty re: tech is planetary in scope, with the powerless ‘AI
ethics’ funding wave as a recent example. The digital behemoth cannot be
allocated. As venture capital is the financial motor driving exponential
hyper-growth, it is a mystery how Big Tech creates monopolies overnight that
become both invisible and untouchable when venture capital is left out of the
equation.
The back cover quote of Nosthoff’s book is crisp and clear: “Cybernetics is
everywhere, like air.” The goal of critical theory is then to make this
invisible ubiquity of the digital visible again. The crisis the Trump 2
administration has thrown the European liberal establishment into will be a
test case if this challenge is taken up, yes or no. Despite calls for ‘digital
sovereignty’, the dominant regressive conservatism is neither pointing in the
direction of critique nor that of alternatives. Understanding stagnation is
what’s on the agenda. This means we need to add this new insight to the
definition: cybernetics as the science of communication, control, and
stabilization in complex systems. Entropy is not just a warning for a possible
collapse; it should be seen as a key part of the cybernetic process.
Nosthoff is critical of Baudrillard’s cynical attitude, yet keeps coming back
to him. This is interesting because Baudrillard cannot be called a critic of
cybernetics. Nonetheless, she brings his basic motif of disappearance into play
(also present in Paul Virilio’s work). She convincingly links these two
Parisians with Günther Anders’ concept of antiquity. The thesis is that notions
such as cybernetic feedback loops only intervene deeply in the hardware,
software, and everyday lives of billions of users after their disappearance.
This is where the decisive effect of cybernetic principles—including their
criticism—lies. According to Anders, humans are beings who must always first
appropriate their world technically. The task is then, in the words of Günther
Anders, “to uncover the fact of concealment itself.” This is the project of the
present critique of ‘cybernetization’—and the core of this large-scale study.
Nosthoff’s goal is neither to save nor to reconstruct cybernetics. She is
concerned with accurately describing the effects of power. This might be why
she did not call for the establishment of ‘Cybernetics 2.0’, as Geneviève Bell
did in 2021 when she founded the School of Cybernetics at the Australian
National University. A few weeks before his untimely death in August 2020,
French philosopher of technology Bernard Stiegler made a similar call for an
informatique théorique (a new theory of computer science). Landed in Rotterdam,
Yuk Hui is working in the same direction. However, the work done here could
certainly be read as the historical and contemporary basis for such an
institute. So far, most of the research conducted in Berlin has been in
journalism, law, and the social sciences. Digital theory and criticism from a
humanities perspective remains small. Her ambitions are clearly different from
the slow bureaucratic politics that solely aim at regulation after the fact. It
is in this void that her “art of digital governance” can come into play. That’s
the visionary note: without a set of critical notions, Nosthoff will not issue
any control instructions. No control without critique. Prisms need to be built
into the design of information systems from the start. This is the right moment
to switch to the interview.
GL: Tell me if I am wrong, but I got a strong sense that your work embodies a
strong will to continue and update the work of Günther Anders. He’s perhaps not
so well known in the Anglo world, but still respected in
philosophy-of-technology circles. Where do you situate the beginnings of this
phenomenal work that you’ve delivered?
AVN: Philosophically, I am indebted to Anders’ work in many respects.
Fortunately, the long-awaited English translation of Anders’s magnum opus, The
Obsolescence of Man, has just been published by the University of Minnesota
Press, thanks to the brilliant efforts of the translators and Anders scholars
Chris Müller and Christian Dries. I hope there’s growing interest in Anders’s
work, and I am positive that interest will only grow given the increased
availability of his work in English. Anders’s thinking seems almost uncannily
timely, which, in my opinion, has to do with his philosophical method of
exaggeration, combined with critical anticipation. Together with Felix, I have
interpreted this method as the intention to think through the totalitarian
potentials and tendencies of cybernetic technologies, and prolong their
potentially destructive tendencies into the future; Anders therefore referred
to himself as a “vorwärts gekehrte Historiker”; an inverted historicist. We
once wrote an article about this aspect and Anders’s contemporary relevance for
Thesis Eleven.
Reading Anders today, I see many observations he already made at the beginning
of the process of cybernetisation regarding the non-neutrality of technologies
when they are no longer conceived as mere means. Or those regarding the
difficulty that moral problems are often posed, and discussed only after the
(technological) fact, i.e., when a technological normalization and
standardization process has already, as it were, colonized the life worlds of
subjects, to borrow Habermas’s expression. Anders is probably best known for
his concept of “Promethean shame”, by which he means the chasm between the
perfect machine and the subject who feels ashamed when confronted with it
because of his or her felt or alleged limitations. I could never really make
much sense of this concept until Chat GPT was introduced, as I have the strong
intuition, and I can confirm this from conversations I have with students, that
many of them, especially at the beginning of their studies, feel they lack the
ability to meet the level of (alleged) perfection when they compare their own
writing with a chatbot’s generative content. Reversing this feeling of
Promethean shame is central to teaching in the age of Large Language Models.
My work is inspired by Anders’ positions. His oeuvre continues to challenge us,
as subjects living in a digital age. Even more, it now confronts us in an age
that is pervaded by so-called “genAI,” where several of its concepts are easily
attributable to phenomena that, in such concretion, it was unable to really
anticipate. That’s precisely why, in retrospect, I would say it lends much
credence to Anders’s philosophical method.
GL: Another unknown critic of cybernetics is the German philosopher Hans Jonas.
His article “Critique of Cybernetics” is from 1953. You compare his piece with
Margaret Mead’s later remarks. Can you explain what these early reflections on
this emerging philosophy of computing were about?
AVN: It was fascinating to me to reread Jonas’s early critique of cybernetics.
And a coincidence that I read it roughly at the same time as I read several
early writings that emerged at the beginning of second-order cybernetics,
Mead’s (and later on Heinz von Foerster’s) writings as central pieces. Of
particular importance to me were Mead’s remarks made before the term
“second-order cybernetics” was coined by Heinz von Foerster. It is important to
explain Mead’s theses first before highlighting the centrality of Jonas’s
earlier critique: Mead argues that cybernetics needs to self-transform to
develop. This was at a time when cybernetics faced an existential crisis.
First, cybernetics had become so broad a term that it was unclear what it
actually meant, making the prior “glamour field” seem rather vague. Second,
cybernetics entered a moment of legitimation crisis in the West as soon as it
became relevant to real-existing socialist contexts, especially the GDR and the
USSR. Mead made this claim during a conference hosted by the Society for
General Systems Theory, which was close to cybernetics. An organization needs
to constantly reflect on itself to adapt to its environment and survive, and
she reiterated this proposition in the context of the Conference Proceedings of
a Symposium of the American Society for Cybernetics. This is where the
second-order perspective is taking shape, and Heinz von Foerster then edits the
text and coins the term “second-order cybernetics” to describe the
cybernetisation of cybernetics. Von Foerster then later claimed that he came to
understand that the problem with first-order-cybernetics was that it, for
instance––thinking of the external observer perspective in which the mind was
described in the sense of neural nets, think of McCulloch and Pitts, for
instance, or Ashby’s homeostat––that first order cybernetics essentially could
not come up with a way to account for its own descriptions: How can a thinking
subject explain the processes of thinking when it is thinking? This is where
von Foerster identifies first-order cybernetics as a blind spot and proposes
recursivity.
What’s fascinating is that Jonas makes similar statements when he criticizes
first-order cybernetics in his early critique. Absurdly enough, he does not
only describe the context in which a group of cyberneticists at a conference of
cybernetics were to describe itself––i.e., the context in which
second-order-cybernetics would then effectively emerge from many years
later––but more importantly, he criticizes that cybernetics is unable to
account for itself, since if the human mind, for instance, were practically
reducible to binary digits and neural signal transmission the cybernetician was
unable to account for how he can even reach that explanation in thinking. I did
some archival work to find out whether Jonas’s text might have been read by
cyberneticists at the time, given the obvious similarities. Bertalanffy was,
interestingly enough, in contact with Jonas, and he was also an important
figure in the emergence of second-order cybernetics. But I could not prove that
Jonas’s text was actually read by the key figures of cybernetics.
It remains interesting that it took first-order cybernetics so long to respond
to a quite similar critique. Or rather, to develop it itself. This has a lot to
do with a general affect against philosophical investigation and critique. This
affect is very visible in early cybernetics––although cybernetics claimed to be
interdisciplinary and a new, open field, it was almost stubborn when it came to
philosophical accounts that, for instance, questioned their techniques of
analogization between human and machinic behaviour, or between the workings of
the mind and the brain; or the way in which they came to ascribe intentionality
to both humans and machines, or how they used systemic and abstract terminology
to describe social, technical, or biological systems, thus erasing all
(material) differences among them. Instead of engaging with such critiques,
cybernetics simply ignored them or deemed them irrelevant.
GL: Your approach is author-centrist. In your study, the critique of systems is
not voiced by engineers, users, programmers, activists, artists, or
entrepreneurs for that matter, but by philosophers. How did they come to this
knowledge? This is a sincere question. If we follow Stafford Beer’s “purpose of
a system is what it does” and map these “doings” in the light of (self)
organization and bottom-up planning, how relevant is it to reflect on the
system as a whole? The Gen Z-driven Hegel memes indicate a need to go for the
whole, a techno-totality. Is the choice between detailed case studies and big
ideas a false one?
AVN: What’s of great importance is the dialectical method that takes into
account the interrelations and contradictions between the whole and the part,
the abstract and the particular. It’s always a challenge to find this position
between closeness to the phenomenon and the distance of the Begriff, Adorno
would say, Begriff und Sache, without operating in violent abstractions. In
Beer’s maxim, I see the problem that systems that center solely on operability
and self-reproduction lose sight of any purpose, that’s their apoliticality,
which I find deeply problematic, as it is easily instrumentalized politically.
That’s close to the circularity we have around generative AI and the ideology
of total scalability, rather than using a particular dataset to train a
particular model for a democratic end. Ironically, in the context of Cybersyn,
Beer kept track of both––mostly at least––he rarely lost sight of the political
purpose of what he tried to accomplish there, using cybernetics. That’s where I
still see the potential of his approach.
Regarding my work and the way it reads philosophical positions, this is why I
focused so much on authors: I did not ascribe to a concrete philosophical
viewpoint, as I did not want to hypostatize. That is also why I chose a
genealogical method, both to describe and understand the development of
cybernetics, and the development of critiques of cybernetics–– and the
intertwinements between both. I do reflect on the normative critiques made at
the time, and uttered, for instance, against the new terminology of systems
that, as Habermas and Wolf-Dieter Narr argued, was too abstract and too far
removed from social phenomena to account for. As they argued, the language was
completely detached from political processes or from political questions
concerning inequality, participation, and resistance. Importantly, some of
these critiques emerged at a time when cybernetics and its central premises
were still subject to debate. This was at a time when there was a distance
between cybernetics and its critics; between the phenomenon in question and the
critical theories that discussed it; between, as it were, the object and the
subject, more broadly.
The landscape changed with the advent of cybernetic capitalism and the personal
computer, when control tools seemed to have been ‘decentralized’ and dismantled
from the technocratic control of the state. Critics, all of a sudden, seemed to
be somewhat part of, or in the midst of, what they criticized. And that was
very consciously reflected upon, as a central dilemma that provoked the
question of where and how to find a place from where to observe critically, and
find a distance to the object, by each of the theorists I engaged with, from
Anders to Habermas to Baudrillard and Tiqqun––as is well known, Deleuze and
Guattari borrowed the term “plane” from Bateson, which shows how they
integrated cybernetic thinking whilst at the same time attempting to transcend
its circularity. Each critic found a different answer. Anders formulated a
philosophy of critical exaggeration in the context of his idea of a moral
phantasy (moralische Phantasie). Habermas sought to limit the
systemic-technological colonization of the lifeworld through communicative
reasoning. Baudrillard used the figure of the Möbius strip to describe the
clashing of binaries in the course of what could be termed a digital
binarization in the midst of his somewhat ironic, provocative mode of critique.
And Tiqqun utilized strategic tactics of anti-net resistance, as it were–forms
of noise production, interruptions. The external observer problem, as
problematic by both second-order cybernetics and poststructuralism, is not
easily attainable nowadays. But this does not mean we have to throw a
critically self-reflective (not self-circular!) objectivity or a process of
critical enlightenment (in Foucault’s sense of interrupting the hegemonic forms
of being-governed) away, as Heinz von Foerster at points seemingly does, who to
my account throws away conceptuality as such for some lose account of
“cybernethics”, as he terms it, which, however, appears random to me at points.
When Adorno claims that there is objectivity in suffering and that this is the
condition of all truth, this focus can attain and preserve a particular form of
ethical objectivity that can survive the clash of the subject-object
distinction. My problem with second-order cybernetics is that it starts from
the system and returns to it in a recursive, not reflective, form. The system
comes first, always–– whereas in an ethics of alterity perspective, what is
named the Other, the Third, or the “Nichtidentische”, with Adorno (who, to my
account, shares some affinities with an ethics of alterity), plays an entirely
different role. This is where ethics starts: precisely from some place that
transcends the system, from some Other that remains unknowable. Cybernetic and
post-structuralism are therefore, to my account, very distinct from each other,
also with regard to their thinking of futurity.
GL: Cybernetics, as the art of controlling the context, failed. Overload is
systemic. As Stiegler warns, the danger of entropy is imminent. Can we still
‘steer’ and adapt to these new circumstances?
AVN: What we currently observe is close to your argument that we are
approaching the extinction internet moment, where networks, the internet as
such, and ‘smart’ systems are simply not working anymore, where they produce
their moments of “enshittification”, bullshit or slop, where they reproduce
racism, systemically, and operate on steroids, where they reproduce and
intensify an overload of extraction. AI hallucinations and Elon Musk’s recent
construction of 35 illegal gas turbines in Memphis to power his xAI are the
most obvious examples of the system’s irrationality. We’re approaching a
tipping point, not just with regard to the climate catastrophe but also with
regard to the evolution of the internet. Both these dynamics intersect.
Regarding the history of cybernetics, adaptation narratives began to change
when the term was no longer solely associated with stabilization, as it was in
the early days of the Cold War, as its Greek origin “kybernetes” suggests
(meaning “steersman”). Although back then it was, of course, also connected to
entropy and moments of crisis, it was always concerned with creating order from
noise and diminishing chaos. Yet, during the dotcom bubble and the early
instabilities of the new economy, it became clear that cybernetics was not
necessarily a technique of stabilization but rather produced its own
instabilities; this is why Virilio spoke of “integral accidents”. This moment
coincides with the becoming hegemonic of the second-order paradigm, as well as
with the advent of complexity, chaos, and systems theory.
What we face is a series of attempts to solve ongoing politico-economic crises
with adaptation techniques. These, however, not only provoke further problems
of adaptation (think of climate engineering “techno-fix” strategies) but also
reproduce the systemic status quo without really addressing the causes of
systemic injustices, inequalities, etc. Technology in this context is never a
solution to a problem; it is always a (cheap, and only supposed) solution and a
problem. The question then, if we take seriously the fact that solutionism is
manufacturing problems, is how to find a better way to address the adaptation
and steering problems of the present. The response here should be to not frame
political problems as adaptation problems: we have, for a long time, made this
mistake. Take nudging, which was prevalent in the aftermath of the 2008
financial crisis. Nudging is a neo-cybernetic technique of governing and
adapting that ignores the systemic causes of the crises we have faced and still
face today, positing the illusion of a post-ideological world. It is no
coincidence that nudging was widely used in smart city initiatives, in which
citizens were sold “smartness” and related concepts. Instead, we need to
politicize narratives of cybernetics and adaptation. Cybernetics is connected
to a post-political illusion, one that describes or creates systems in such a
way that they are–– allegedly–– technological alternatives to politics. But the
problem lies precisely in the depoliticization of adaptation.
GL: I am glad you ‘limited’ your study to the role of critique in the overall
roll-out of ‘cybernetization’ of society. This is arguably only one side of the
story. The other is the split in 1955 and the separation between the
feedback-networking side and the artificial intelligence direction, led by
Minsky and others. You did very well in keeping the ‘knowledge’ part of machine
learning out of your investigations. Given the current hype around large
language models and AI, how do you defend your choice? How do you relate to the
current AI onslaught?
AVN: There is a connection between the different traditions and lineages of
cybernetics and cybernetisation and what is currently termed AI or generative
AI. First, Wiener’s work in the context of the so-called
Anti-Aircraft-Predictor develops an early form of predictive analytics on the
grounds of anticipations of the future that are based on the collection of past
data on behavior, in his context, regarding the directions of enemy aircraft.
This technique remains relevant for LLMs that rely on probabilistic assessments
and stochastic methods. Also, we should not forget that McCulloch and Pitts
developed the first neural net at the beginning of the 1940s, with the aim of
simulating the brain’s behavior; similar to Ashby’s construction of the
homeostat, a model of a living brain. There is a link between early
cybernetics’ accounts of the brain-computer-nexus and notions of AGI–so I was
not at all surprised, for instance, when Marc Andreessen, the
right-wing-libertarian author of the now infamous “Techno-Optimist Manifesto,”
quotes McCulloch and Pitts’s seminal text from 1943 to claim that the
“revolution” of AGI builds on McCulloch’s and Pitts’s proposition. This is
entirely superficial, but I still think it is interesting that he refers to
them. What Andreessen means here is: In the early forties, the idea was to
imitate the human brain, and we have forgotten that, later on, with narrow AI
and now with projects aiming for AGI, superintelligence, etc. We are back to
building that. Is it a coincidence that Anthropic’s Claude is named after
Claude Shannon? The Shannon, whose mathematical information theory was
instrumental to cybernetics, especially to Wiener’s and Deutsch’s work? I don’t
think so.
GL: The last part of your study describes the past ‘internet’ decades, defined
by omnipresent connectivity, wearables, and platforms. One element I would like
to take out here is ‘the social’. You discuss Alex Pentland’s Social Physics
and Tiqqun’s Cybernetic Hypothesis. The way algorithms work on social media is
how billions experience the feedback mechanism today. It is hard to imagine how
social relations in the near future will not be shaped by code. How do you
envision ‘the social’ as part of a form of governance that is transparent and
not extractive?
AVN: I would probably not call it transparent in the first place. Rather, we
have to think of the social beyond social media, and beyond mere connectivity,
and what we would need, among other things, is a different imaginary based on a
different terminology. I always felt close to an ethics of alterity, in a
Derridean, maybe also Levinasian sense. Levinas thinks responsibility in the
sense of being able to respond. It is a response to a demand placed on me by
the Other, and this demand is asymmetrical: it places the Other first, which is
important, as it goes far beyond a mere network-based approach to (supposed)
decentrality and horizontality. Response-ability understood in this sense can
never be reduced to feedback. Yet, it is equally important not to stop here.
For Levinas, the ethical relationleads to politics through the inevitable
integration of a “Third”. The political, understood in this sense, is a
disruption of the ethical relation, but it is necessary to offer the
possibility for justice in being; in that sense, it is also related to the
inevitably political tasks of organization and conceptualization.
The question for me would be of how to think through a response-able politics
in this sense, for the current age in which the social as a category is
corroded, hollowed out, by the bad sociality and the schlechte Unendlichkeit of
social media. Such a response-able politics can be close to your (and Ned
Rossiter’s) approach to “organized networks”: creating relations that are not
reducible to mere connectivity, giving subjects the space to evolve beyond
being reduced to nodes in a network. Next to the pragmatic question of
organization (and politics, for that matter), we have to pay equal attention to
the question of relationality and sociality: Is there a way to reimagine how we
relate to each other that goes beyond connectivity?
GL: According to Chris Anderson, the plenitude of data provokes the “end of
theory”. From now on, the program is the theory. Together with Felix and
others, you are running the Critical Data Lab. How do you see data in relation
to theory? Should philosophers indeed know more than just prompting and have
classic programming skills? How do you envision a critical tech education?
German institutions are notoriously conservative and behind in this respect.
Should philosophy still be taught the way Adorno and Arendt once gave their
seminars? What are your early experiences in Oldenburg in this respect?
AWN: Tech critique today needs to pay close attention to the developments
regarding tech-fascism. This should be the first task, and these issues are
also at the heart of the seminars I am teaching, often together with Felix.
What we try to make clear is that we need to investigate the manifold forms of
power involved in techno-authoritarianism–from communicative power (think of
Musk’s propaganda on X), to geopolitical power (such as Musk’s usage of
Starlink in the context of the Ukrainian war), and financial power (to point to
the obvious). Next is to come up with multiple ways of resisting these forms of
power. I am sympathetic to the now-emerging AI refusal movements around the
globe and to the revival of critique in the Luddite tradition (which I do not
view as a refusal of tech tout court, though). It is important that we give the
protesters currently protesting the construction of new data centers a voice,
and that we also address our media coverage of their tactics. I would like to
point here to Karen Hao’s list of organizations and movements. Take the Cables
of Resistance Conference in Berlin. It is important to understand that a whole
new movement that politicizes tech critically is currently underway.
It is equally important that we correct the misconception that digitization is
an immaterial process by examining the materialities of the digital. Many
current forms of so-called AI are regressive, not progressive; ineffective, not
effective; centralizing, not decentralizing; closing, not opening. On that
note, responding to these powers requires different forms of critique:
deconstructionist techniques, ideological critique, and discourse analysis can
help unravel the tactics and strategies of Big Tech actors, for instance. But
methods from computational social sciences can help here as well, for instance,
in identifying new ideological narratives in the context of communities such as
Effective Altruism and the TESCREAL bundle. Regarding EA, narratives about AI
have, for instance, shifted to “superintelligence” narratives around the time
Bostrom’s book of the same name was published a few years ago, prompting EA to
shift to longtermism, as Carolin Müller (from ZeMKI in Bremen) recently argued.
Several of these narrative nuances have shaped AI regulations, so it is vital
to understand which discursive dynamics are at play. So, yes, (critical) data
can also play a role in informing critique, I would say, also with regard to
making things visible–think of projects such as Data in Feminicide in this
respect. I think here we need to be creative in developing new methods as well.
I am fortunate in Oldenburg to be teaching bright, excellent students who are
well-read in Kant, Hegel, and Marx, as well as in post-structuralism. Many of
them are deeply concerned about contemporary developments regarding the rise of
techno-authoritarianism. They are engaged and open to critical thinking through
contemporary phenomena surrounding the digital condition together. Many of my
seminars are very experimental; recently, I used a very open format, which I
co-taught with my colleague Tilman Hannemann, a scholar of religious studies,
to examine the religious aspects of techno-fascism. We invited Carolin Müller,
who conducts ethnographic research in EA circles, to present her work. The idea
for the seminar emerged from the shared conviction that we can learn from each
other’s perspectives and use trans-disciplinary approaches and methods to
better understand the complexities of particular phenomena.
GL: An element of future theory could be the regression of cybernetics up to
the point where civic computing collapses and returns to its military origins.
Back to Friedrich Kittler? You use the term ‘authoritarian cybernetics’ for
this. Do you see a need to move on from all the platform blues and address the
more urgent issues of our time, like ‘antifa’ resistance against AfD and the
militarization of society after the Ukraine invasion by Russia? In my recent
work, I have pointed out the violent turn in internet culture. Is the subtle
term ‘governmentality’ still appropriate in such a dire situation? Can we still
state that power has disappeared into invisible infrastructures? All these
contradictory developments seem to accelerate and collide at the same time.
AVN: I am afraid we are already advancing well into this stage. To my account,
it is no coincidence that Palantir is openly speaking of the new era of
deterrence, whereas this new deterrence will be, as Alex Karp is very explicit
about in his propaganda, AI-based. Several of the major themes of first-order
cybernetics recur currently with regard to AI weapons systems: predicting the
future to anticipate the enemy’s moves, etc. In a promotional video, Palantir
recently proclaimed that they help address the challenge of “navigating the fog
of war,” which directly alludes to the navigability metaphor central to
cybernetic and neo-cybernetic steering techniques. The centrality of
cybernetics becomes visible, too, when Musk currently draws on cybernetics
implicitly when he speaks of “cyber-powers” in relation to his company
Neuralink and the improvement of brain-computer interfaces, when he dubs people
as “always already cyborgs”, or when he speaks of X as a “cybernetic collective
superintelligence”.
Equally, Zuckerberg speaks of companies as “learning systems”––which is a
central figure of second-order cybernetics––while Andreessen directly alludes
to cybernetics in his Techno-Optimist-Manifesto. This is not mere rhetoric and
is not only reflective of the influence cybernetics has had on the
counterculture that then shaped cyberculture, which, in turn, shaped Silicon
Valley. To my mind, there is more to that; in fact, I do think that we are
witnessing the reemergence of central motifs that were prevalent, especially in
first-order cybernetics. Although we also witness a fusion of both forms of
cybernetics, I would say: Think of the way in which Musk uses X, of course, He
uses it as a quite centralized form of propaganda, in a sense, but this form of
propaganda is still using feedback, also horizontal feedback, for messages and
content to spread. So centralized modes of steering and governing are more
prevalent in current forms of cybernetics, but modes of circularity and
self-recursion remain part of the whole picture. In my account, this fusion of
centralized and decentralized forms of control is central to cybernetic
authoritarianism.
GL: You may know Made in China, Designed in California, Criticized in Europe.
How do you respond to this tragic role of Europe that has reduced its role to
the legal work of the world’s tech regulator? These days, Berlin is not exactly
perceived as the global capital of free thinking. Is Europe tired and
provincial? How can we envision a vital culture of criticism that is not bitter
and does not arrive a decade too late with remarks on the side? Is Germany
still the right place to develop a philosophy of technology for the 21st
century?
AVN: A philosophy of technology for the 21st century would most certainly have
to involve the perspectives of those who are mostly affected by the negative
impact of current digital systems. This is a perspective that I am trying to
hint at as well at the end of my book: That we need to make visible what is
usually hidden from view with regard to cybernetic capitalism, and that
includes of course the exploitative labor that is the basis of current AI
systems, and the ecological costs of LLMs, for instance, including the
communities that are mostly affected by them.
We won’t solve these structural issues with ethics committees. You are right to
point out that our legal systems are simply not fast enough or well-equipped
enough to adequately regulate these technologies. A ban on big tech lobbying
might be worth considering as an additional measure in this context, given the
extent of lobbying efforts in recent years and their influence on AI
regulations. We have to come up with ways to democratize these systems from the
ground up, a way of democratization that includes the full stack. Your concept
of stacktivism is to the point here, and I am also thinking of people who have
recently worked on the Euro Stack framework. One task is also to bring together
the various groups currently mobilizing against AI. If we come up with ways to
democratically govern these technologies, the question of regulation is less
relevant–logically, as these systems would be designed democratically, from the
ground up. It is important to emphasize that these democratic structures
already exist, think of the Fediverse. Not all here is optimal, but this can
serve as a basis for the course. Yet, to arrive at a more democratic digital
society, we also need to cultivate a different sociotechnical imaginary when
using and speaking about these technologies.
—
Interview with links:
https://networkcultures.org/geert/2026/06/13/interview-with-anna-verena-nosthoff/.
Info on her German book:
https://www.suhrkamp.de/buch/anna-verena-nosthoff-kybernetik-und-kritik-t-9783518300794.
