TY - ABST

T1 - Data (sets) for the future?

T2 - Media Futures

AU - Wiehn, Tanja

PY - 2023/6/15

Y1 - 2023/6/15

N2 - Synthetic data is an emerging field in data science. This growing interest stems from well-known issues with data from real-world systems: gender and racial bias in data sets, data shortage and concerns about privacy. Unlike organic data, synthetic data is generated in algorithmic models, characterized as an alternative to manually collected and processed data. It can be created as a novel output by systems of artificial intelligence, such as GANs (General Adversarial Networks) or generated in fully simulated digital worlds (Nikolenko, 2021; Steinhoff, 2022). In light of the call for papers, this contribution engages with the emerging field of synthetic data as a powerful technology of media futures. The central questions of this paper are: How will the emergence of synthetic data affect the politics of future imaginaries? How can the distinction between synthetic and organic data theoretically be characterized? The empirical work of this paper is based on fieldwork and expert interviews with data scientists. Theoretically, the paper draws on critical data studies, feminist theory and critique of techno- determinism (Amoore 2020; D’Ignazio & Klein 2020; Thylstrup 2022). Synthetic data is praised as cost-effective and nearly indistinguishable from organic data. Dangers of synthetic data in the form of deep fakes are already at the forefront of everyday media usage (Meikle 2022). Moreover, it holds the promise to complete data sets for the sake of better AI models. The building of synthetic data sets in artistic research practices demonstrates cases for its purpose for social justice (“VFrame”, Harvey 2022). The paper thus aims to engage with a nuanced discussion of synthetic data’s impact of media futures to re-think binaries of dystopia and utopia, as well as synthetic and organic.

AB - Synthetic data is an emerging field in data science. This growing interest stems from well-known issues with data from real-world systems: gender and racial bias in data sets, data shortage and concerns about privacy. Unlike organic data, synthetic data is generated in algorithmic models, characterized as an alternative to manually collected and processed data. It can be created as a novel output by systems of artificial intelligence, such as GANs (General Adversarial Networks) or generated in fully simulated digital worlds (Nikolenko, 2021; Steinhoff, 2022). In light of the call for papers, this contribution engages with the emerging field of synthetic data as a powerful technology of media futures. The central questions of this paper are: How will the emergence of synthetic data affect the politics of future imaginaries? How can the distinction between synthetic and organic data theoretically be characterized? The empirical work of this paper is based on fieldwork and expert interviews with data scientists. Theoretically, the paper draws on critical data studies, feminist theory and critique of techno- determinism (Amoore 2020; D’Ignazio & Klein 2020; Thylstrup 2022). Synthetic data is praised as cost-effective and nearly indistinguishable from organic data. Dangers of synthetic data in the form of deep fakes are already at the forefront of everyday media usage (Meikle 2022). Moreover, it holds the promise to complete data sets for the sake of better AI models. The building of synthetic data sets in artistic research practices demonstrates cases for its purpose for social justice (“VFrame”, Harvey 2022). The paper thus aims to engage with a nuanced discussion of synthetic data’s impact of media futures to re-think binaries of dystopia and utopia, as well as synthetic and organic.

M3 - Conference abstract for conference

Y2 - 15 June 2023 through 16 June 2023

ER -