What do you think?
Rate this book
If...Then: Algorithmic Power and Politics
We live in a world in which Google's search algorithms determine how we access information, Facebook's News Feed algorithms shape how we socialize, and Netflix collaborative filtering algorithms choose the media products we consume. As such, we live algorithmic lives. Life, however, is not blindly controlled or determined by algorithms. Nor are we simply victims of an ever-expanding artificial intelligence. Rather than looking at how technologies shape or are shaped by political institutions, this book is concerned with the ways in which informational infrastructure may be considered political in its capacity to shape social and cultural life. It looks specifically at the conditions of algorithmic life -- how algorithms work, both materially and discursively, to create the conditions for sociality and connectivity. The book argues that the most important aspect of algorithms is not what they are in terms of their specific technical details but rather how they become part of
social practices and how different people enlist them as powerful brokers of information, communication and society. If we truly want to engage with the promises of automation and predictive analytics entailed by the promises of "big data", we also need to understand the contours of algorithmic life that condition such practices. Setting out to explore both the specific uses of algorithms and the cultural forms they generate, this book offers a novel understanding of the power and politics of algorithmic life as grounded in case studies that explore the material-discursive dimensions of software.
social practices and how different people enlist them as powerful brokers of information, communication and society. If we truly want to engage with the promises of automation and predictive analytics entailed by the promises of "big data", we also need to understand the contours of algorithmic life that condition such practices. Setting out to explore both the specific uses of algorithms and the cultural forms they generate, this book offers a novel understanding of the power and politics of algorithmic life as grounded in case studies that explore the material-discursive dimensions of software.
216 pages, Paperback
Published June 26, 2018
24 people are currently reading
195 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 13 of 13 reviews
May 24, 2020
As a computer science major, I thought I knew algorithms. But Prof. Bucher's sociopolitical lens showed me how wrong I was, particularly as machine learning becomes ubiquitous. In this regard, the book gets 5 stars. It, however, reads like a dissertation with very little rhetoric to keep the reader's attention. This isn't too surprising considering the notes and bibliography take up more than a quarter of the book itself.
If you can trudge through the text, however, the core tenets of If...Then are enlightening. Despite what companies and politicians may say, algorithms are not neutral and they're not unknowable black boxes. An algorithm isn't just the written code, but the data it's trained on, the users that feed it more data, the perceptions users have of the "algorithm" and how this perception changes their actions, the worldviews of the engineers and what they believe is "normal."
The list goes on and on, and still isn't exhaustive, which is Bucher's entire point.
And a bit of a side-note: her phrase "fear of invisibility" is a beautifully apt description of what social media's algorithms are designed to bring out in its users. It flips the classic Panopticon example on its head; instead of instilling fear of perpetual surveillance to keep others disciplined, it instills a fear that others can't see you until you self-discipline according to the algorithm's desires.
If you can trudge through the text, however, the core tenets of If...Then are enlightening. Despite what companies and politicians may say, algorithms are not neutral and they're not unknowable black boxes. An algorithm isn't just the written code, but the data it's trained on, the users that feed it more data, the perceptions users have of the "algorithm" and how this perception changes their actions, the worldviews of the engineers and what they believe is "normal."
The list goes on and on, and still isn't exhaustive, which is Bucher's entire point.
And a bit of a side-note: her phrase "fear of invisibility" is a beautifully apt description of what social media's algorithms are designed to bring out in its users. It flips the classic Panopticon example on its head; instead of instilling fear of perpetual surveillance to keep others disciplined, it instills a fear that others can't see you until you self-discipline according to the algorithm's desires.
September 30, 2019
The book does a good job of outlining how algorithms are used by the main social websites and attempts to create a background framework in theory, mainly Foucault’s power (‘power is everywhere’).
I thought it was an interesting read, but I would’ve been interested also in the author’s views on how to act in this setting and maybe some predictions.
Also I thought a lot of the algorithms seem to be created as an evolutionary process (e.g through continuous A/B testing), thus becoming ‘fitter’ over time.
I thought it was an interesting read, but I would’ve been interested also in the author’s views on how to act in this setting and maybe some predictions.
Also I thought a lot of the algorithms seem to be created as an evolutionary process (e.g through continuous A/B testing), thus becoming ‘fitter’ over time.
Read
July 26, 2024I read this book because it is beginning to constitute somewhat of a classic in one of my scholarly fields of interest, namely critical data/algorithm studies. Hence, my reading of the book has largely been with an eye to what it contributes to this field of inquiry; or, the insights it synthesizes. More specifically, I have been interested primarily in the way Bucher conceptualizes algorithms as well as her methodological approach. The more empirically-oriented chapters (ch. 4-6) seemed interesting, but I didn't bother to read them in any great depth. What follows is a sketch of the individual chapter's general arguments.
The first chapter sets out to clarify the book's overall aim which is to "consider the power and politics of software and algorithms that condense and construct the conditions for the intelligible and sensible in our current media environment" (p. 3). The chapter's guiding concept is that of "programmed sociality" which serves to sensitize us to the fact that sociality - humans' social activities - are incresingly shaped by programs of all kinds. As an example, Facebook promotes a particular kind of sociality which is programmed through "likes", "friends", and further concepts proper to the way in which Facebook' programs sociality. Her point is that Facebook re-defines what friendship means in this day and age, contrasting FB's enactment of friendship with traditional philosophical perspectives on it )p. 13). Interestingly, Bucher also outlines her concept of politics as drawing from Science and Technology Studies (STS), specifically from post-ANT analyses of ontological politics; her interest, hence, is in analyzing how algorithms help bring about certain realities rather than others. Moreover, she draws from a decidedly Foucauldian perspective on power (p. 3) as relational and productive: power is not in possession, it is differentially distributed; power is not merely negative/prohibitive, it also generates things. She argues that algorithms are "evolving, dynamic, and relational processes hinging on a complex set of actors, both humans and nonhumans" (p. 14). She also mobilizes Simondon's concept of "transduction" (p. 14) to argue that Facebook constantly participates in the reshaping of what friendship means today. Moreover, she mobilizes his notion of technicity: "algorithms and software, in this view, do not determine what friendships are in any absolute or fixed sense. Rather, technicity usefully emphasizes the ways in which algorithms are entities that fundamentally hinge on people's practices and interaction" (p. 14).
The second chapter, "The Multiplicity of Algorithms", provides an interesting and complex conceptualization of algorithms. Before that, however, she provides a brief genealogy of algorithms. Importantly, algorithms always work in concert: "to be actually operational, algorithms work in tandem not only with data structures but also with a whole assemblage of elements, including data types, databses, compilers, hardware, CPU, and so forth" (p. 22). She then goes on to explain how machine learning upends previous forms of algorithms. In this day and age, importantly, algorithms can shape events: "algorithms have the ability performatively to change the way events unfold, or, at the very least, change their interpretation" (p. 28). She further draws on Langdon Winner's (1980) famous paper to suggest that algorithms are never neutral. Moreover, she is interested in the ontological politics of algorithms: "What is at stake [...] in addressing the ontological politics of algorithms is not so much an understanding of what exactly the algorithm is or the moments in which it acts [...] but, rather, those moments in which they are enacted and made to matter as part of specific contexts and situations" (p. 40). In sum, according to her it makes most sense to stay away from abstractions and attend to specific contexts of algorithms shaping practices.
The third chapter, "Neither Black nor Box", was perhaps the most interest one. She begins by problematizing the famous metaphor of algorithms as "black boxes" by tracing a genealogy of black boxes. She suggests, instead of accepting the black box metaphor, to ask what kind of purposes this metaphor serves in the first place. For example, it effectively leads to a mystification and essentialization of algorithms that invites a methodological excavation of the "core" of algorithms; however, a processual and relational perspective on algorithms clearly has its issues with such an approach. Concretely, for instance, to look into the source code of a given algorithm might only be so helpful in better understanding its workings. How have algorithms come to be seen as black boxes in the first place and who benefits from this metaphor? What are the longer genealogies that link algorithms to the black box metaphor? Methodologically, Bucher suggests to "un-know" (p. 46) algorithms: "making the familiar seem slightly more unfamiliar" (p. 46). Then, she goes on to theorize algorithms as multiple, processual, and heterogeneous (p. 48); often what is called "the" XY algorithm is an amalgamation of multiple smaller ones acting in concert (p. 48). Again she criticizes the black box metaphor: "algorithms in contemporary media platforms are neither black nor box but eventful. For a conceptualization of algorithms, this implies a rejection of essences and permanence and an ontological shift towards a world of process and relations" (p. 48). She further draws on Whitehead to foreground questions around what algorithms do rather than what they are (p. 49). Relatedly, she argues that the agency of algorithms resides not within them but is distributed across the elements of the assemblages within which they operate (p. 51). Accordingly, her provocation is also to ask "When is an algorithm?" (p. 55). She wants us to attend to agency in specific situations, rather than presume agency to reside within the confines of human minds.
I largely skipped chatper 4.
Chapter five, "Affective Landscapes" contains an interesting theorization of everyday encounters with algorithms (p. 93). She foregrounds the "micropolitics" (p. 94) of algorithms as emanating from their affective and phenomenological qualities. Moreover, she here engages the concept of imaginaries to conceptualize these encounters.
In chapter 6, "Programming The News", I stumbled across a very interesting mobilization of Lucy Suchman's work who calls for "the need to attend to the boundary work through which entities are defined" (p. 126). Moreover, she suggests that algorithms produce problematizations (p. 142); more precisely, algorithms re-shape the very definition of journalism.
The conclusion brings together her conceptual and empirical work. Crucially: "How algorithms come to matter in contemporary society is not about trying to define what they are or at what points they act, but rather about questioning the ways in which they are enacted, and come together to make different versions of reality" (p. 152). Revisiting Simondon's notion of technicity, she suggests that algorithms' "capacity to produce sociality always already occurs in relation to other elements, and as part of an assemblage through which these elements take on their meaning in the first place" (p. 153). She closes very poignantly: "In looking at what constitutes an algorithmic life, the central question therefore becomes: who or what gets to be part of whatever is being articulated as the algorithm" (p. 159).
Bucher provides a very interesting perspective on the power and politics of algorithms. She effectively challenges essentialist accounts by emphasizing the processual, multiple, and heterogeneous nature of algorithms. She does this in a very sophisticated way, weaving together the scholarship of philosophers such as Whitehead and Simondon as well as STS scholars such as Suchman, Latour, and Barad. If you are well-read in these fields, the book won't provide many new insights. Nonetheless, it constitues a very readable and fun synthesis of these crucial insights into algorithms.
The first chapter sets out to clarify the book's overall aim which is to "consider the power and politics of software and algorithms that condense and construct the conditions for the intelligible and sensible in our current media environment" (p. 3). The chapter's guiding concept is that of "programmed sociality" which serves to sensitize us to the fact that sociality - humans' social activities - are incresingly shaped by programs of all kinds. As an example, Facebook promotes a particular kind of sociality which is programmed through "likes", "friends", and further concepts proper to the way in which Facebook' programs sociality. Her point is that Facebook re-defines what friendship means in this day and age, contrasting FB's enactment of friendship with traditional philosophical perspectives on it )p. 13). Interestingly, Bucher also outlines her concept of politics as drawing from Science and Technology Studies (STS), specifically from post-ANT analyses of ontological politics; her interest, hence, is in analyzing how algorithms help bring about certain realities rather than others. Moreover, she draws from a decidedly Foucauldian perspective on power (p. 3) as relational and productive: power is not in possession, it is differentially distributed; power is not merely negative/prohibitive, it also generates things. She argues that algorithms are "evolving, dynamic, and relational processes hinging on a complex set of actors, both humans and nonhumans" (p. 14). She also mobilizes Simondon's concept of "transduction" (p. 14) to argue that Facebook constantly participates in the reshaping of what friendship means today. Moreover, she mobilizes his notion of technicity: "algorithms and software, in this view, do not determine what friendships are in any absolute or fixed sense. Rather, technicity usefully emphasizes the ways in which algorithms are entities that fundamentally hinge on people's practices and interaction" (p. 14).
The second chapter, "The Multiplicity of Algorithms", provides an interesting and complex conceptualization of algorithms. Before that, however, she provides a brief genealogy of algorithms. Importantly, algorithms always work in concert: "to be actually operational, algorithms work in tandem not only with data structures but also with a whole assemblage of elements, including data types, databses, compilers, hardware, CPU, and so forth" (p. 22). She then goes on to explain how machine learning upends previous forms of algorithms. In this day and age, importantly, algorithms can shape events: "algorithms have the ability performatively to change the way events unfold, or, at the very least, change their interpretation" (p. 28). She further draws on Langdon Winner's (1980) famous paper to suggest that algorithms are never neutral. Moreover, she is interested in the ontological politics of algorithms: "What is at stake [...] in addressing the ontological politics of algorithms is not so much an understanding of what exactly the algorithm is or the moments in which it acts [...] but, rather, those moments in which they are enacted and made to matter as part of specific contexts and situations" (p. 40). In sum, according to her it makes most sense to stay away from abstractions and attend to specific contexts of algorithms shaping practices.
The third chapter, "Neither Black nor Box", was perhaps the most interest one. She begins by problematizing the famous metaphor of algorithms as "black boxes" by tracing a genealogy of black boxes. She suggests, instead of accepting the black box metaphor, to ask what kind of purposes this metaphor serves in the first place. For example, it effectively leads to a mystification and essentialization of algorithms that invites a methodological excavation of the "core" of algorithms; however, a processual and relational perspective on algorithms clearly has its issues with such an approach. Concretely, for instance, to look into the source code of a given algorithm might only be so helpful in better understanding its workings. How have algorithms come to be seen as black boxes in the first place and who benefits from this metaphor? What are the longer genealogies that link algorithms to the black box metaphor? Methodologically, Bucher suggests to "un-know" (p. 46) algorithms: "making the familiar seem slightly more unfamiliar" (p. 46). Then, she goes on to theorize algorithms as multiple, processual, and heterogeneous (p. 48); often what is called "the" XY algorithm is an amalgamation of multiple smaller ones acting in concert (p. 48). Again she criticizes the black box metaphor: "algorithms in contemporary media platforms are neither black nor box but eventful. For a conceptualization of algorithms, this implies a rejection of essences and permanence and an ontological shift towards a world of process and relations" (p. 48). She further draws on Whitehead to foreground questions around what algorithms do rather than what they are (p. 49). Relatedly, she argues that the agency of algorithms resides not within them but is distributed across the elements of the assemblages within which they operate (p. 51). Accordingly, her provocation is also to ask "When is an algorithm?" (p. 55). She wants us to attend to agency in specific situations, rather than presume agency to reside within the confines of human minds.
I largely skipped chatper 4.
Chapter five, "Affective Landscapes" contains an interesting theorization of everyday encounters with algorithms (p. 93). She foregrounds the "micropolitics" (p. 94) of algorithms as emanating from their affective and phenomenological qualities. Moreover, she here engages the concept of imaginaries to conceptualize these encounters.
In chapter 6, "Programming The News", I stumbled across a very interesting mobilization of Lucy Suchman's work who calls for "the need to attend to the boundary work through which entities are defined" (p. 126). Moreover, she suggests that algorithms produce problematizations (p. 142); more precisely, algorithms re-shape the very definition of journalism.
The conclusion brings together her conceptual and empirical work. Crucially: "How algorithms come to matter in contemporary society is not about trying to define what they are or at what points they act, but rather about questioning the ways in which they are enacted, and come together to make different versions of reality" (p. 152). Revisiting Simondon's notion of technicity, she suggests that algorithms' "capacity to produce sociality always already occurs in relation to other elements, and as part of an assemblage through which these elements take on their meaning in the first place" (p. 153). She closes very poignantly: "In looking at what constitutes an algorithmic life, the central question therefore becomes: who or what gets to be part of whatever is being articulated as the algorithm" (p. 159).
Bucher provides a very interesting perspective on the power and politics of algorithms. She effectively challenges essentialist accounts by emphasizing the processual, multiple, and heterogeneous nature of algorithms. She does this in a very sophisticated way, weaving together the scholarship of philosophers such as Whitehead and Simondon as well as STS scholars such as Suchman, Latour, and Barad. If you are well-read in these fields, the book won't provide many new insights. Nonetheless, it constitues a very readable and fun synthesis of these crucial insights into algorithms.
Read
August 12, 2020An idea-rich and interesting social science theory book on algorithms.
First chapter: explains how social media platforms, notably Facebook, deliberately cultured a new typology of friendship, where friendship has become a 'programmed sociality' primed in every way to improve Facebook's revenue. That is, lonely or unengaged users are less likely to revisit their Facebook account; Facebook has an inherent interest to linking people to other active nodes in the network, such as friends or active pages, as a 'hook' to get them to commit to regular visits. Furthermore, the 'friends' that are most worthwhile to Facebook have either high activity (update regularly) or high affinity (most likely to share similar ad preferences as you), and these people are likely to rise in one's newsfeed. Since Facebook profits from accurately understanding- and selling-- profiles arising from these 'friendships', Facebook subtly engenders these online relationships to be computable, quantifiable, and procedural, so that its algorithms can understand and manipulate these relationships as nodes on a network.
Second chapter: technical definitions of algorithms
Third chapter: a sustained exhortation not to presume the popular metaphor of algorithms as 'black boxes', because this premise defeats attempts to understand algorithms. This provides algorithm creators with prepared extenuation from culpability. Rather, we should understand that agency for algorithms is distributed. Furthermore, the author proposes three methodologies to unpack the black box of algorithms: reverse-engineering, phenomenological encounters, and interrogating their configurations.
Fourth chapter: How Facebook culture works, how Facebook newsfeeds wor, the threat of selective invisibility afflicting some subjects or participants of the Facebook newsfeed
First chapter: explains how social media platforms, notably Facebook, deliberately cultured a new typology of friendship, where friendship has become a 'programmed sociality' primed in every way to improve Facebook's revenue. That is, lonely or unengaged users are less likely to revisit their Facebook account; Facebook has an inherent interest to linking people to other active nodes in the network, such as friends or active pages, as a 'hook' to get them to commit to regular visits. Furthermore, the 'friends' that are most worthwhile to Facebook have either high activity (update regularly) or high affinity (most likely to share similar ad preferences as you), and these people are likely to rise in one's newsfeed. Since Facebook profits from accurately understanding- and selling-- profiles arising from these 'friendships', Facebook subtly engenders these online relationships to be computable, quantifiable, and procedural, so that its algorithms can understand and manipulate these relationships as nodes on a network.
Second chapter: technical definitions of algorithms
Third chapter: a sustained exhortation not to presume the popular metaphor of algorithms as 'black boxes', because this premise defeats attempts to understand algorithms. This provides algorithm creators with prepared extenuation from culpability. Rather, we should understand that agency for algorithms is distributed. Furthermore, the author proposes three methodologies to unpack the black box of algorithms: reverse-engineering, phenomenological encounters, and interrogating their configurations.
Fourth chapter: How Facebook culture works, how Facebook newsfeeds wor, the threat of selective invisibility afflicting some subjects or participants of the Facebook newsfeed
January 1, 2019
an embarrassment
March 7, 2022
A very interesting entry point into the literature on algorithms and science and technology studies, which offers productive ways to avoid the pitfalls of the "black box" metaphor.
August 5, 2024
If...Then is an academic text focusing on artificial intelligence algorithms in modern day, and how these algorithms affect and effect our interactions. It is separated roughly into two sections: chapters 1-3 focus on the theoretical groundwork for Bucher's view of algorithms, while chapters 4-7 focus on applications and conclusions. The theoretical section highlights the artificiality created by algorithms, given Facebook "friends" as a significant example of this artificiality. It also covers both the ontology of algorithms and their epistemology - in other words, the essence of the algorithms and the knowledge required to engage with them. The take-aways here are that we need to focus no on the "what" of algorithms, but on the "why" and the "how", and to do this requires unknowledge - a Science and Technology Studies term which regards shifting and changing our perceptions towards a topic and the ways in which we think we know it. The application chapters take these theoretical developments and focus them in on three issues: the visibility of algorithms (and the invisibility created by turning to a panopticon-style algorithm), the micropolitics of mundane and everyday data and experience influenced by algorithms, and finally the ways in which algorithms affect reliability (which is done by comparing Scandinavian "editor controlled algorithms" with the so-called "California code" of Facebook, Twitter, and Instagram).
While there were some salvageable portions in the theoretical section, and my rating of 2 stars is defensible solely by reference to these salvageable portions, on the whole I was frustrated by this book. For full disclosure, my job involves curating an academic research portfolio - my research is on safe and ethical artificial intelligence in healthcare. Many of the arguments made in the theoretical section suffer from what I personally consider a major flaw of the STS approach to research: it's all telling and no showing, with many fancy words and terms that don't map to the world outside the academy. And I say that as someone with a graduate degree in Philosophy, so I am also come from a field that traditionally misses the point of application to the real world! With a little more care, however, many of the arguments could be viable and indeed influential in understanding the effects of algorithms. It just wasn't done well in what was presented.
Secondly, and more frustratingly, the research methodology in the applied section was awful. In one of the chapters Bucher's data is "based on data collected in the time period of April-August 2011, using my own personal Facebook profile" but fails to mention any biases this could cause (especially since the chapter was focusing on the influence of algorithmic timelines rather than live feed timelines!) nor ways to mitigate undue influence of the researcher's implicit biases. Likewise, another chapter is described as having the data come from 35 Twitter users she personally chose to reach out to, and then a secondary study came from the friend list of only one of these 35 users. Social science and hard science research have requirements for adequate research standards, and these were not met in this book. We aren't told whether Bucher's qualitative data met saturation, how sample sizes were selected, what the inclusion or exclusion criteria were, how biases were managed, etc.
While there were some salvageable portions in the theoretical section, and my rating of 2 stars is defensible solely by reference to these salvageable portions, on the whole I was frustrated by this book. For full disclosure, my job involves curating an academic research portfolio - my research is on safe and ethical artificial intelligence in healthcare. Many of the arguments made in the theoretical section suffer from what I personally consider a major flaw of the STS approach to research: it's all telling and no showing, with many fancy words and terms that don't map to the world outside the academy. And I say that as someone with a graduate degree in Philosophy, so I am also come from a field that traditionally misses the point of application to the real world! With a little more care, however, many of the arguments could be viable and indeed influential in understanding the effects of algorithms. It just wasn't done well in what was presented.
Secondly, and more frustratingly, the research methodology in the applied section was awful. In one of the chapters Bucher's data is "based on data collected in the time period of April-August 2011, using my own personal Facebook profile" but fails to mention any biases this could cause (especially since the chapter was focusing on the influence of algorithmic timelines rather than live feed timelines!) nor ways to mitigate undue influence of the researcher's implicit biases. Likewise, another chapter is described as having the data come from 35 Twitter users she personally chose to reach out to, and then a secondary study came from the friend list of only one of these 35 users. Social science and hard science research have requirements for adequate research standards, and these were not met in this book. We aren't told whether Bucher's qualitative data met saturation, how sample sizes were selected, what the inclusion or exclusion criteria were, how biases were managed, etc.
January 11, 2024
A Bucher é sempre um alento para quem estuda algoritmos e plataformas. Sua perspectiva contingente e nada determinista nos ajuda a aterrar quando refletimos sobre o caráter performativo dos algoritmos. Importante leitura que nos ajuda a construir insights poderosos sobre o que ela chama de "sociabilidade programada", sobre hibridismos e fronteiras. Afinal, estamos todos entrelaçados nessas redes, mas cada entrelaçamento é diferente. E é importante nos atentarmos para isso.
March 26, 2021
If you are already well-read in the critical algorithm studies, this book will not add much but review what you have already seen. However, if you are a beginner, this may help you get a good grasp of recent literature.
December 3, 2019
I wish it thought more about practice of resistance besides the bits near end and in conclusion, but theory is on point. A ton of good insights.
September 22, 2022
Incredibly dense. She said so many things, but somehow didn’t say anything at all.
May 18, 2023
Very Technical
Good book but very technical!! There was a lot of good information in this one. It’s interesting to see how algorithms actually work.
Good book but very technical!! There was a lot of good information in this one. It’s interesting to see how algorithms actually work.
July 23, 2025
内容其实也比较新的，关于推荐算法、算法如何反映和放大现实中既定的偏见、参与者如何与算法互动等等，选择的视角（社交、新闻）也都是我比较熟悉且感兴趣的。但是读到最后“我们要负起责任”的这个观点似乎又让我有点不满意，当然“我”们作为用户要知道“这与我相关”、并且积极捍卫自己的权利，但更多时候用户面对算法的时候依旧是弱势，少数群体更是，更应该负起责任的是掌握知识和权威的人。
Displaying 1 - 13 of 13 reviews