Books about so-called “AI”, the hype around “AI” and the connection between “AI” and authoritarian and fascist forces are much needed these days. They are important resources for communities to understand, organise against and resist “AI”. Aline and I have referenced some of them in our writing on this blog, among them Dan McQuillan’s “Resisting AI” and Jathan Sadowski’s “The Mechanic And The Luddite“. When I engage with the topic of “AI”, people often start talking about their own recommendations for books or talks which is always very welcome. Among these recommendations is a new book by German professor of philosophy Rainer Mühlhoff, titled “KI und der neue Faschismus” (“AI and the new fascism”). Mühlhoff also spoke about the issue on many occasions, e.g. together with Aline at this year’s re:publica in Berlin. He is a well respected academic who has published extensively about “AI”.
This blog article will attempt to unpack how several arguments and general topics of the book are examples of what can be described as “convenience criticism”. It presents a form of criticism that is easily digestible, doesn’t require the reader to take a position and is largely agreeable also to those in power. This kind of criticism limits itself to (correct) observations of symptoms and delivers critical descriptions instead of substantial systemic analysis and critique. Other examples of “convenience criticism” is white-liberal “feminist” literature that rightfully calls out inequality but solves it by advocating for women to join men in the roles of oppressors. Works about poverty that deny the systematic and intentional function of impoverishment in our society and works about neo-colonialism that praise Europe’s alleged exemplary human rights commitments are further examples. Convenience criticism usually describes symptoms accurately but refrains from addressing the structural causes in favour of responses that merely mitigate some symptoms. To be clear: There is nothing wrong with being approachable or avoiding provocation. Good analysis is not defined by how many people are upset by it. And sometimes analysis is also just incomplete simply due to lack of time or space or other resources. What I am saying is: Convenience criticism is not necessarily intentional. But it is harmful. It takes up space that is much needed for actions that actually address the root causes. It dampens efforts to engage in meaningful resistance, protest and change.
In the case of Mühlhoff’s book I regret to say that it is part of convenience criticism to a large extent. It uses the vocabulary and references of actual criticism while legitimising and stabilising systemic causes of “AI” becoming the threat to society that it is. I feel that the book is negatively impacting our resistance against “AI” and fascism in a way that warrants a closer look. To do so, I will focus on three shortcomings: The book avoids questioning the economic dynamics that shape “AI” and its enabling role for fascism. It also plays into the narrative of the “democratic centre” and disqualifies necessary paths of actions as “radical” and “undemocratic”. It eventually resorts to solutions that have proven themselves inadequate to challenge the status quo or the role of “AI” in it.
Did someone say capitalism?
One notable gap in the book is the lack of structural analysis of capitalism as an economic and social system of control and power. This is striking as the book mentions the word “capitalism” numerous times. It mentions capitalist motives, venture capitalists, concentration of market power and capitalist production of value. It also follows the popular trend of talking about capitalism mostly in connection with certain qualifiers like “digital” or “Silicon Valley“, suggesting that it is the specific type of capitalism one should take aim with, not capitalism in general. This kind of language is a key part of convenience criticism. Mühlhoff’s book uses certain words without proper context and analysis and gives the misleading impression of including an analysis of capitalism in the broader topic of “AI” and fascism. What it actually achieves is the opposite. Mühlhoff refers to capitalism as a natural and unchallenged background phenomenon and thereby removes it from the scope of the book. Jathan Sadowski (The Mechanic And The Luddite, 2025) rightfully warned against using modifiers to the word “capitalism” in such a way: “Modifying the word capitalism with adjectives like rentier or platform or surveillance is useful for focusing on certain features of capitalism; […]. But these modifiers can also obscure our analysis if they cause us to mistake one part for the whole problem or to overlook how the future is connected to the past. Suddenly your critique slips into a nostalgic plea for the halcyon days of good capitalism before the thing isolated by the specific modifier ruined everything else.” A famous example of this kind of nostalgia for good capitalism is Shoshanna Zuboff’s book “Surveillance capitalism” from 2019. Evgeny Morozov already wrote a must-read critique of Zuboff’s book and its negative impact on the digital policy discourse. In his review of the book for The Baffler, he explains how Zuboff’s analysis comes at a price: A “sense of confusion with regard to the origins, operations, and vulnerabilities of digital capitalism.“ This is what Mühlhoff’s book does as well. His readers are left with the impression that capitalism is not to blame for the state of today’s misery, but the misguided or misinformed actions of certain companies and policy-makers are. Mühlhoff’s book leaves its audience with a vague rejection of abuse of market power while vindicating the existence of markets and the power they inevitably create for the owners of corporations and resources.
This, in turn, leads to another gap: The book misses explaining how the inequality and power hierarchies that define capitalism on a systemic level are closely linked to the rise of fascism in Europe of the 1930s. Capitalism is a deeply political system to organise production and access to goods. The ideology behind capitalism, economic liberalism, calls for free markets, autonomy of property owners and commoditised human labour. This ideology, as Ishay Landa (The Apprentice’s Sorcerer – Liberal Tradition and Fascism, 2012) and Quinn Slobodian (Hayek’s Bastards, 2025) have shown, was and is closely linked to the ideas of early right-wing and fascist thinkers in 1920 and 1930 Germany, the alt-right in today’s United States or to the rise of Germany’s new fascist party, the AfD. The link was so clear that Max Horkheimer famously said before World War 2: “Whoever is not willing to talk about capitalism should also keep quiet about fascism.” Mühlhoff’s book references Horkheimer, but it does not reference this specific quote. Instead, it treats fascism as an ideology foreign to our own social order. Fascism – according to Mühlhoff’s book – seems to be the result of a flawed eccentric elite with too much power and bad ideologies. Everything was “more or less fine”, he writes echoing Zuboff, “while we could trust Silicon Valley to adhere to Google’s slogan »Don’t be evil«”. That Google now plays nice with right-wing regimes is – to him – an unfortunate bow to Trump’s alt-right regime, not the result of cold economic necessities that require corporations to adjust to new political conditions. Consequently, “AI” and its destructive effects are not linked to economic logic, but are discussed as detached from them.
Mühlhoff’s book correctly describes many of “AI” dangers but shies away from connecting “AI” to the fundamental workings of our capitalist society. Such a connection would demand space for questions about alternatives to capitalism and for the – unpopular – conclusion that “AI” may not be socially beneficial as long as it is ruled by capitalist logic or – maybe – not be beneficial at all. Such questions and considerations are well beyond the scope set by the boundaries of convenience criticism. The book however still claims to do just that. In the conclusion, Mühlhoff writes that “we need to study the profit and power motives that drive our technology in order to gain the necessary critical instruments to responsibly talk about the AI future”. It would have been great had the book provided more analysis of that kind or acknowledged its lack in that regard.
The myth of the democratic center
Another shortcoming is closely linked to this gap of an analysis of “AI”’s link to capitalism. Mühlhoff often explicitly positions his analysis within the current constitutional order, within its legal system and his perspective as part of the “democratic centre”. His guidelines and goal posts are law and order, proper procedure and good liberal statesmanship. The dangers to liberal democracy are not coming from within but from external forces: from tech ideologies and ultra-right forces. Theirs are the “wrong hands” that technology might end up in. Our hands, of course, are the right ones. These ultra-right forces, he writes, can meaningfully only be countered by the “forces of the centre”, by more debate and by more educational work. Again, this wording is symptomatic of convenience criticism. The message is clear: We are the good guys, we just need to get it together, do the right thing and defeat the forces of evil that appeared at our doorstep out of nowhere. The consideration that the very constitutional and legal system of European nations is a part of the problem has little room in Mühlhoff’s analysis. The idea that we are fighting an evil of our own making and that we will fight it until we fundamentally change how our society is organised is cast out into the realm of radicalism.
This is where Mühlhoff most visibly leaves the role of an observer and takes a clear position: When he talks about “the state”. The role of the state is – according to him – to “contain power imbalances that threaten the economy or our society”. He is then quick to limit this role through emphasising that the state must refrain from “reappropriating AI technology and its commercial infrastructures”. Such reappropriation – Mühlhoff writes – may be conceivable only when “accompanied by careful regulation, public and democratic governance and working institutional separation of powers”. This, however, still seems too dangerous to Mühlhoff. He warns that this would play in the hands of right-wing movements who would “merge state power and technological power”. This hesitation to advocate in favour of a public and democratic governance because of the risk of authoritarian takeover strongly echoes cyberlibertarian dogma. Cyberlibertarian thinking generally distrusts collective control over technology as potentially authoritarian, as David Golumbia explained in his 2024 book “Cyberlibertarianism“. While the danger of authoritarian takeover surely is very real, this kind of argument works as a knockout argument. If taken to its logical conclusion, it would mean that all collectively governed infrastructures are sleeping risks for fascist takeover and it would encourage us to quickly privatise what little public infrastructure we have left. This conclusion – while surely not intended – would culminate in a radical market ideology which reverts to a struggle of all against all and survival of the fittest. This kind of ideology nourishes other ideas of superiority and is the reason why capitalism and fascism were so closely linked in the past. By not addressing this link, Mühlhoff’s book unfortunately achieves the opposite of what it intended to do. It creates gaps that make the book easier to swallow but also easier to refer to by the very forces he tries to weaken. This is why convenience criticism is so dangerous.
The master’s tools
The book culminates in a call for regulation. What we need – according to Mühlhoff – are rules that cushion some of the inequalities that are amplified by “AI”. We need conscious regulation that shapes our digital future. And we should be “proud” of Europe and its achievements in tech regulation, the GDPR being an explicit example Mühlhoff calls upon. However, the GDPR with all its achievement is still a regulation that makes it perfectly legal to commercialise data, to compete for profits in a data economy and to engage in a race to the bottom of exploitative business practices. The GDPR states that it aims at enabling the “free flow of data” within the unified European market. The current symptoms of this can be seen in the consent-or-pay-models of online publishers, behavioural advertising and algorithmic amplification of harmful online content. These aren’t failures in an otherwise healthy system. These aren’t perversions of an otherwise fair economic order. These are parts of the normal operations of capitalism, as Sadowski explains. No legal framework that remains within the boundaries of capitalism is able to escape it, as Marie Thøgersen (Ruling The Cloud, 2025) has shown in the context of international cybersecurity law.
In the end, it is the disparity between claim and substance that makes the book such an unfortunate example of convenience criticism. There are valid reasons to analyse problems with a limited scope. Many books do so, dissecting the “hype” aspect or the ideologies behind so-called “artificial general intelligence”. These approaches would still benefit from including thorough structural analysis about the socio-economic dynamics that lead to today’s situation (and some even include just that despite their more focussed subjects). But leaving out systemic analysis is less of a problematic omission in these cases. When a book – as it is the case here – tackles such a broad topic like the interplay between “AI” and fascism, leaving out key systemic issues is a different story. In the beginning of his book, Mühlhoff writes that some recommended not to use the word “fascism” in the title of his “otherwise very much necessary critical analysis of the developments in the USA and – to a lesser degree – in Europe”. He then reassures his readers that such an omission would be a “mistake” and writes that it is of great importance to understand current developments through the methods and properties of “fascism”. He then even defines economic motives and “occasionally purely capitalist motivated support” as a major aspect of the rise of fascism. By setting the stage with such ambition and yet being noticeably quiet about major systemic issues behind the links between capitalism, “AI” and fascism, he narrows the horizon of possibilities for fighting fascism. Audre Lorde famously said that the master’s tools will never dismantle the master’s house. A critical analysis of the political and economic conditions that lead to today’s alliance between “AI” and fascism should include a description of what the master’s house is made of and would encourage thinking about other tools to dismantle it. Mühlhoff’s book, in contrast, fails to inspire such considerations and discourages truly transformative approaches.
Foto von Zoshua Colah auf Unsplash