Resisting Artificial Intelligence
Notes from Resisting AI: an anti-fascist approach to artificial intelligence by Dan McQuillan.
These are my notes taken from this book (published by Bristol University Press in 2022). Dan McQuillan is a lecturer in creative and social computing at Goldsmiths (London). I have known Dan since the late 1980s and we have been active in various campaigning and political groups.
I encourage people to buy Dan’s book.
“AI represents a technological shift in the framework of society that will amplify austerity while enabling authoritarian politics… The net effect of applied AI is to amplify existing inequalities and injustices, deepening existing divisions on the way to full-on algorithmic authoritarianism…
“As we look forward with trepidation to the consequences of the climate crisis, with the likelihood that privilege will be defended, responsibility deflected and the vulnerable sacrificed, our priority for advanced technologies should be to ask not only how they can be prevented from intensifying harm but how we can reassert the primacy of the common good…
“While machine learning has some clever mathematical tricks up its sleeve, it’s important to grasp that it is a brute force mathematical process. There’s no actual intelligence in artificial intelligence… One of the most important aspects of machine learning is not that it heralds the sudden spark of consciousness in silicon but that it is a set of computational methods with political implications…
“Whatever we think of specific AI applications, accepting AI means we are implicitly signing up for an environment of pervasive data surveillance and centralised control… What this opens up is opportunities for unaccountable decisions, unjust exclusions and exploitative speculations…
“Deep learning is a complex set of nested mathematical operations that are off the scale in terms of anything we can grasp directly…the landscape rarely consists of a unique valley, and can be filled with various dips and crevices that can trap an unwary algorithm. At the very least, this invisible complexity should cast doubt on any claim made by deep learning to produce a singular truth…
“If artificial intelligence has a soundtrack, it’s the deafening whir of cooling fans in the server farms…
“Google’s self-evident hypocrisy and its subsequent gaslighting and undermining of former staff highlights the gulf between ethical statements and actual practice… Ethics, as mobilised by corporates and their fellow travellers, is not only a matter of fine sounding statements but a method of depoliticising and individualising problems that really should be a matter of collective political concern…
“AI isn’t aiming for understanding: it only cares about what will come next, not why… In general, applied algorithms are performative in that they help to reshape the very phenomenon they are supposedly modelling…however sophisticated or creative AI might seem to be, its modelling is stuck in abstractions drawn from the past, and so becomes a rearrangement of the way things have been rather than an imagining of the way things could be… This innate conservatism makes AI a good fit with the broader tendency in contemporary society known as ‘tech solutionism’ – the substitution of advanced technology for any serious attempt to address the structural causes of a problem…
“Capitalist Realism refers to a ‘widespread sense that not only is capitalism the only viable political and economic system, but also that it is now impossible even to imagine a coherent alternative to it’…
“Applying AI’s abstractions to the institutional status quo has the effect of intensifying structural violence… AI tries to emulate science by revealing a hidden mathematical order in the world that is superior to our direct experience…the operations of AI depend on datafication: the presentation of the world as data separated from the continuity of experience…
“Converting the social world into data for the benefit of AI is to convert ourselves into a standing reserve for optimisation and prediction. We are abstracted and reduced to that which be usefully optimised…
“The operations of AI make it a good fit for neoliberalism’s retreat from social care and unrelenting hostility to organised labour. While AI is a technology that claims to calculate risk and therefore reduce uncertainty, it actually acts to increase precarity. Applied AI is not so much a means of prediction as an engine of precaritisation…
“AI also amplifies precariousness on an ecological level. Data centres increasingly consume scarce water resources in regions already impacted by global warming while the water ‘becomes a repository for electronic waste and derivative toxins, making toxicity a feature of surrounding systems and ecologies’…
“Platform work, online or offline, comes without the protections, such as sick pay, holiday entitlement, pensions or health and safety, that were hard won by the historical struggles of organised labour… The fear of a privileged algorithmic observer leads to an anxious performance of compliance…
“Rather than heralding an alternative sci-fi future, AI can be more plausibly understood as an upgrade to the existing bureaucratic order… Under these conditions, AI will produce the kind of thoughtlessness that Hannah Arendt warned us about. Thoughtlessness manifests as the inability to critique instructions, the lack of reflection on consequences and a commitment to the belief that the correct ordering is being carried out…Thoughtlessness enables participants to evade any responsibility for wider harms…
“While AI is heralded as a futuristic form of productive technology that will bring abundance for all, its methods of helping to decide who gets what, when and how are actually forms of rationing. Under austerity, AI becomes machinery for the reproduction of scarcity…
“The idea that ‘intelligence’ is a single entity that can be abstracted to a single number has been thoroughly critiqued… The general assumption of Artificial General Intelligence (AGI) believers is that mind is the same as intelligence, which is itself understood as logic and rationality. A commitment to AGI and the associated reifications of rationalism often comes with social imaginaries that revolve around intellectual elitism and beliefs about innate and biologised superiority…
“There are certainly strong overlaps between the core operational idea of optimisation and a eugenicist perspective on populations… If there’s one thing that history teaches us, it’s that we need to be very wary of where the systematic application of discriminative ordering can end up… Techno-authoritarians sneer at democracy as an outdated operating system which they can replace with their own blend of autocracy and algorithms…
“While AI is a genuinely novel approach to computation, what it offers in terms of social application is a reactionary intensification of existing hierarchies. Likewise, fascism offers the image and experience of revolution without fundamentally altering the relations of production or property ownership. AI is technosocial solutionism while fascism is ultranationalistic solutionism…
“How we can resist, interrupt and replace AI – given the entrenched momentum behind systems that produce algorithmic states of exception, this is going to take more than appeals to ethics or tinkering with regulation…
“Current AI overlooks the work of care that underpins the world and replaces it with datafied models of reality that are disconnected and domineering… Rather than relying on the traditional scientific virtues of certainty and neutrality, the key axes of post-normal science become ‘uncertainty’ and ‘values’…
“The argument of this book is that AI is best understood in the same way: not as a system that measures and represents the world, but as an apparatus that helps to produce aspects of the world through the specific exclusions it sets up… According to Niels Bohr, the way you measure – the specific set up of the apparatus – materially affects what you find…
“The shock to the system delivered by Covid-19 made it clear that problems can be reinvented and solved beyond the boundaries of the previously possible. When there’s a pandemic, it becomes suddenly possible to house the homeless, it becomes possible to provide forms of universal basic income, it becomes possible to have roads that are not filled with choking traffic…
“AI is the steam hammer of limited imagination, a solution to problems defined in administrative offices and enforced through predictive boundaries…
“We need to connect our matters of care to a politics of change…we draw on the politics of mutual aid and solidarity to articulate an alternative…aims to generate a different kind of autonomy through workers’ and people’s councils… The mobilisation of mutual aid is a direct counter to algorithmic segregation and carelessness…
“Where the algorithmic delivery of care is scarcified, commodotised and individualised, mutual aid is expansive, anti-discriminatory and collectivised… Where mutual aid is the means to tackle a shared need, solidarity is the basis for the struggle against the systems that create scarcity…
“To contest these new enclosures we can invoke the commons, not only in the traditional form of shared natural and cultural resources, but encompassing forms of organising, relating and acting… Commoning is both a refusal of segregation and an assertion of the common good…
“The Luddite movement of 1811-16 faced an uncanny similar set of dilemmas to contemporary society… Although they are best known for their machine breaking, the idea that the Luddites were anti-technology per se is historical disinformation… It was not only their economic situation but their dignity and agency that the new machinery threatened to devastate… AI systems are taking on aspects of what the Luddites would call ‘obnoxious machines’…
“However clever these AI systems appear to be at making recommendations based on data, they will always fail on a social level because they will never recommend liberatory social change…
“Technology is welcome where it supports and extends the commonality, where it acts as part of an apparatus for collective wellbeing… Current AI is an engine of un-commoning, while the goal of an anti-fascist approach to AI is the construction of an apparatus for the expansion of the commons…
“The exponential scaling of computing power is complicit with the economic ideology of unconstrained growth. Perhaps we already have all the computing that we need, and the future is more about recycling, salvaging and repurposing. A free society is its own performance of unfolding complexity, whether or not elements of that rest on digital computation…”
