The collection of vast amounts of data is necessary to the functioning of AI and machine learning based systems. Where that data is personal data, the idea of consent, informed consent, and the redundancy of consent have become a part of debates on technology and rights. Governments across the world are looking to the potential of AI to open new markets and drive economic growth. In 2018, the Government of India, through Niti Aayog (formerly the Planning Commission), released a discussion paper titled'National Strategy for Artificial Intelligence.' This document stated that "for accelerated adoption of a highly collaborative technology like AI, the government has to play the critical role of a catalyst in supporting partnerships, providing access to infrastructure, fostering innovation through research and creating the demand by seeking solutions for addressing various governmental needs."

Recent conversations around gender and AI have centred around the need to understand gender beyond the binary of male and female. For example, facial recognition technology used by Uber in the US has problems with correctly recognising transgender persons (see here and here). Yet Uber is no exception. The U.S. National Science Foundation, for example, has highlighted research that shows that "facial analysis services performed consistently worse on transgender individuals, and were universally unable to classify non-binary genders." According to CNN Business,¹ "The way a computer sees gender isn't always the same way people see it. A growing number of terms for describing one's gender are becoming common in everyday life."

For decades, social researchers have argued that there is much to be learned when things go wrong.¹ In this essay, we explore what can be learned about algorithms when things do not go as anticipated, and propose the concept of algorithm trouble to capture how everyday encounters with artificial intelligence might manifest, at interfaces with users, as unexpected, failing, or wrong events. The word trouble designates a problem, but also a state of confusion and distress. We see algorithm troubles as failures, computer errors, "bugs," but also as unsettling events that may elicit, or even provoke, other perspectives on what it means to live with algorithms -- including through different ways in which these troubles are experienced, as sources of suffering, injustice, humour, or aesthetic experimentation (Meunier et al., 2019). In mapping how problems are produced, the expression algorithm trouble calls attention to what is involved in algorithms beyond computational processes.

AI/ML models are often exported. For example, large tech companies tend to congregate in particular parts of the world and sell software-as-a-service, platform-as-a-service, even surveillance-as-a-service in neatly-bound packages on a subscription basis to individuals, companies, and authorities around the world [1]. As a service/product/labour, AI/ML systems are also frequently exceptionalized where sleek models are intentionally portrayed to magically appear from thin air, skipping the commodity chain altogether. Software and virtual products are often decoupled from their material entanglements -- divorced from the vast lithium farms of the Atacama Desert, the cold data centers underneath the Alps, the data annotation centers scattered across the world, and the digital graveyards in the Korle Lagoon [2], [3]. As a feature of our capitalist society, all global supply chains have hidden components -- whether it is the obfuscation of sweatshops that operate on child labor or the efforts made towards washing the blood off of the diamond industry.

Who is human, what is a machine, and who gets to decide where the boundaries lie?¹ In the field of AI research, who is included and who is excluded in the category of the human? The answer depends on whose knowledges, practices, and modes of living inform your analysis. The field of AI has been shaped significantly by its co-evolution with humanism and the ways that it is entangled with Western Enlightenment thinking, which assert that science and technology -- not, for example, art, religion and mysticism -- are the unquestioned drivers of innovation, linear human progress, and modernity. From the embedding of computing into the human body (e.g., Elon Musk's Neuralink) to the current obsession with (big) data-driven modeling to the media frenzy around Jeff Bezos' Blue Origin rocket launch -- this basic assumption is found in businesses, universities, and public policy.

Companies, governments and institutions in the Global North -- and even in Westernized spaces in the Global South -- often communicate and think about AI through singular lenses that convey over-the-top optimisms or pessimisms about the future. AI is going to save humanity, as promised by many AI corporations.¹ Alternatively, AI is going to destroy us, with big tech corporations and personalities regularly selling us this fear and the idea that their technological innovation will save us from AI or at least prepare us from AI taking over our lives. In the process of conveying these lenses, minoritized communities are further otherized from technoscientific processes; seen as either incapable of having any answers to everyday living problems that may require technological solutions, or seen as passive consumers of technology and capital exploitation. In this essay, I present an alternative discourse of AI creation: one driven by communities, collaboration, and mutual support.

Which languages and language-users are prioritized by digital platforms? Speakers of non-dominant languages are disproportionately subject to algorithmic harms.¹ They confront content moderation algorithms that "only work in certain languages"² on platforms that structurally omit non-Western nations from governance considerations. I call this tendency algolinguicism -- a matrix of automated processes that minoritize language-users outside the Global North and obstruct their access to political participation. This essay addresses digital platforms as sites of algolinguicism.

In late 2020, the presidency of the Council of the European Union (EU) released its conclusions on artificial intelligence and human rights. In contrast to the American and Chinese approaches, this effort formalizes what has been dubbed as "a third way" for artificial intelligence -- tech regulation that prioritizes shared "European values." The EU document focuses on the "underlying idea of human dignity" as the key element of a "human-centric approach to AI." While it states that AI-based solutions can "perpetuate and amplify discrimination, including structural inequalities," it also notes that "one Member State continued to object to the use of the term'gender equality'" in the final draft. This dissenting voice, it was later reported, was that of Poland.

Social scientists amply document algorithmic harms and algorithmic bias, such as discrimination in hiring, medical settings, or the criminal justice system, and for good reason -- these systems produce wide effects and are deployed on a massive scale. Yet there is less attention on how different forms of pleasure, affect, and desire produce and drive both normative and renegade repurposings of these systems. Pleasure, or pleasures, which I take to encompass the expressive life, range of feelings, affective charges routed through technical systems, and the systems of drives that animate social, ecological, and technical worlds, needs to be thought of as an essential, and not always positive, aspect of our technological systems. As AI systems develop and critiques of these systems mount, we will have to come to terms with the following realities: First, forms of desire for control, power, knowledge, and progress gave rise to techno-solutionism in the first place, and repudiating those forms will require cultivating other modes of pleasure. The belief that the problems produced by algorithms can be solved by ever newer forms of technology is deep-seeded and seated in a heady mixture of (white) male cultural norms, ideas about progress that treat those on the receiving end of technological harms as'backward,' and colonial norms that separate out a particular technology from the larger environmental, economic, social, and cultural contexts in which they unfold (Ricaurte 2019, Ullman 2017, Broussard 2018, Forsythe 2001, Heyward-Rotimi 2021).

Regulators in the EU and US have recently drawn attention to the market power and monopolistic behavior of big tech firms. Lawmakers in the EU argue that'traditional businesses are increasingly dependent on a limited number of large online platforms' and that these'gatekeepers' leverage their privileged position to stifle competition and enter new markets at a rapid pace (EPRS 2020). Striking a similar tone, lawmakers in the US claim that these companies have'abused their dominant positions, setting and often dictating prices and rules for commerce, search, advertising, social networking and publishing' (Kang and McCabe 2020). Yet these arguments tend to focus mostly on the conduct and market position of online platforms, while ignoring the underlying technologies by which they operate. But what if techniques like machine learning (ML) are themselves factors in the continuous expansion of already oversized tech giants?