...Sylvie Delacroix Abstract The salient concern, today, is not whether copyright law will “allow robots to learn.” The pressing question is whether the exploitation of the data ecosystem that has made generative AI possible can be made socially sustainable. Just as the human right to water is only...

... for) a seamless and inevitable journey from data to monetizable domain knowledge and useful services. Perhaps the classic application of that critical data-studies framework is Marion Fourcade and Kevin Healy's influential 2017 essay, “Seeing Like a Market,” which advances a comprehensive account of how value...

... , no. 1 ( 1996 ): 44 – 72 . Chun, Wendy Hiu Kyong . 2012 . “ Race and/as Technology; or, How to Do Things to Race .” In Nakamura and Chow-White, Race after the Internet , 7 – 34 . Couldry, Nick , and Ulises A. Mejias . 2019a . The Costs of Connection: How Data Is Colonizing Human...

...Katherine Bode; Lauren M. E. Goodlad Abstract This introductory article calls attention to the shift from the “big data” discourse of the 2000s to the current focus on “AI” in its supposedly “responsible” and “human-centered” forms. Such rhetoric helps deflect attention from the profitable...

...Catherine D'Ignazio Ultimately, in place of “move fast and break things” Marshall proposes “move intentionally and make tech fairer for vulnerable communities” (245) as the slogan of an equity-first data science. This is an urgent message, and this reviewer, for one, hopes that the good humor...

...Sam Lavigne Abstract Web scraping is a technique for automatically downloading and processing web content or converting online text and other media into structured data. This article describes the role that web scraping plays for web businesses and machine learning systems and the fundamental...

..., notice, guidance, and appeal structures for students. [email protected] Copyright © 2024 Kathryn Conrad 2024 privacy data surveillance ethics educators Generative AI 1 fully came to the attention of the public in 2022, first through the coverage of the improvement in large image...

...,” including the data-driven technologies that now claim that name, remains riddled by three core dilemmas: (1) reductive and controversial meanings of “intelligence”; (2) problematic benchmarks and tests for supposedly scientific terms such as “AGI”; and (3) bias, errors, stereotypes, and concentration...

... understandings of language, inaugurated during the imperial age of national print cultures and further consolidated by computing culture itself. However, understanding languages as digital data sets whose patterns can be probabilistically reproduced is a simplistic reduction of an already reified linguistic...

... of the photographic image into a stable and transparent form of data that can be portrayed as a proxy of human vision. Reflecting on the prominence of the photographic snapshot in machine vision discourse, the article traces the path that made this popular cultural practice amenable to the dataset. Following...

... in scale and “data positivism,” researchers adapted these powerful models for the probabilistic scoring of text to chat interaction and other “generative” applications—even though the models generate convincingly humanlike output without any means of tracking its provenance or ensuring its veracity...

... or information access. Their development and deployment as commercial tools in a climate of reductive data positivism and underregulated corporate power overturned a long history in which researchers regarded chatbots as “misaligned” affordances for safe or reliable public use. While the technical underpinnings...

... of these technologies are significant and worth exposing to the public in some detail to engage citizens in future-building. While data technologies of all kinds have always generated waste and enabled violent (extractivist, settler-colonial, capitalist) notions of progress to take hold, there's a renewed, deeply...

... and Maggie Dryden 2024 black box large language models metaphor data justice “GPT-4 Is a Giant Black Box and Its Training Data Remains a Mystery” (Barr 2023 ); “Two-Hundred-Year-Old Math Opens Up AI's Mysterious Black Box” (Choi 2023 ); “ChatGPT, Black Boxes, and Information Dissemination...

... Wasielewski, is therefore a step backward, an argument she unpacks in the following chapters. Chapter 1 outlines how computer vision systems, designed specifically for the analysis of art and built using data sets of disproportionately Western artworks, reify Western ways of seeing. Wasielewski's audit...

... and democratic) way. Existing discussions have been biased by the term itself, which imputes intelligence to the application of procedures like those used in data-driven machine learning (ML). “AI” tells the lay public that these procedures have successfully simulated human thought. The age is now labeled...

... of the dangers of scale in LLMs. Enumerating a variety of problems and harms, these researchers demonstrated how the massive scale of unfiltered data underlay the unreliable and biased condition of their outputs. They recognized the CDO problem with LLMs—that it is low-quality data simply repackaged at scale...

... an adversarial attack on the training data or machine learning process (Oprea, Singhal, and Vassilev 2022 ). More closely related to the environmental analogy would be the case of an entity inadvertently introducing poison into an LLM by training a new AI model on the “synthetic” data produced by earlier models...

... different languages represented in its data pool, dozens of dialects, and “both male and female voice options” (Perez 2022 ). LiveStory, however, is not Khanmigo, the automated learning resource of Khan Academy. Khanmigo, profiled in the Washington Post by staff writer Gillian Brockell ( 2023...

... might hope. There is almost no discussion of successful uses of data-driven machine learning; almost every example provided is of a misuse or failure, which might give the impression that “AI” never succeeds at doing what its proponents claim, when a more judicious survey of examples would show...