Technologies of Speculation:
The limits of knowledge in a data-driven society
(NYU Press, 2020, formerly Fabrications)
Available July 2020. You can now preorder via NYU Press – use the code HONG30 for a 30% discount.
What counts as knowledge in the age of big data and smart machines? In its pursuit of better knowledge, technology is reshaping what counts as knowledge in its own image – and demanding that the rest of us catch up to new machinic standards for what counts as suspicious, informed, employable. In the process, datafication often generates speculation as much as it does information. The push for algorithmic certainty sets loose an expansive array of incomplete archives, speculative judgments and simulated futures where technology meets enduring social and political problems.
Technologies of Speculation traces this technological manufacturing of speculation as knowledge. It shows how unprovable predictions, uncertain data and black-boxed systems are upgraded into the status of fact – with lasting consequences for criminal justice, public opinion, employability, and more. It tells the story of vast dragnet systems constructed to predict the next terrorist, and how familiar forms of prejudice seep into the data by the back door. In software placeholders like ‘Mohammed Badguy’, the fantasy of pure data collides with the old spectre of national purity. It tells the story of smart machines for ubiquitous and automated self-tracking, manufacturing knowledge that paradoxically lies beyond the human senses. Such data is increasingly being taken up by employers, insurers and courts of law, creating imperfect proxies through which my truth can be overruled.
The book situates ongoing controversies over AI and algorithms within a broader societal faith in objective truth and technological progress. It argues that even as datafication leverages this faith to establish its dominance, it is dismantling the longstanding link between knowledge and human reason, rational publics and free individuals. Technologies of Speculation thus emphasises the basic ethical problem underlying contemporary debates over privacy, surveillance and algorithmic bias: who, or what, has the right to the truth of who I am and what is good for me? If data promises objective knowledge, then we must ask in return: knowledge by and for whom, enabling what forms of life for the human subject?