Created: 2025-05-27 10:12
Suchman’s (2023) focus on the “thingness” of AI has deeply resonated with me, not only because it sheds light on the ontological significance of AI, but also because it connects with my own theoretical interests. Suchman argues that AI isn’t a natural or self-evident “thing,” but rather a construct or “figuration.” In popular discourse, AI is often treated as a stable, monolithic entity with agency. However, Suchman points out the inherent contradictions and uncertainties within it.
One key contradiction is that AI, while appearing to be a technological tool, becomes imbued with strong political factors when statistical correlations within datasets are accepted as causal relationships between real-world entities.
The term “thingness” is particularly appealing to me because of my engagement with Actor-Network Theory (ANT) and the new-materialist theories of Knappett and Malafouris (2008) during my doctoral research (more details in Presenting Doctoral Thesis at CSSE 2025). While Suchman doesn’t explicitly state whether his ideas are derived from ANT or new materialism, there are clear connections.
Thingness refers to the inherent, often latent, properties of a material entity that become apparent when it interacts within a network or is acted upon by an agent (Knappett & Malafouris, 2008). Through Suchman’s argument, these inherent properties relate to many scholars’ ontological investigations and interpretations of AI.
For example, Crawford (2021) provides a comprehensive analysis of the various factors behind the concept of AI, including the often-invisible or overlooked human labor. Specifically, Sperling et al. (2022) highlight the significant amount of manual labor by teachers and faculty hidden behind AI’s seemingly automated processes.
Algorithms and data are two other closely connected elements. Bucher (2025) and Noble (2018) discuss algorithms and the political and commercial logic behind them. Bucher argues that algorithms are not neutral or deterministic, but rather evolving and changing. O’Neil (2016) reveals the opacity of algorithms and how big data can exacerbate biases and inequity.
As Bucher (2025) mentions, the popularity of the concept of “AI” overshadows the focus on algorithms. This is precisely what Suchman’s (2023) “thingness” aims to express: a concept that obscures the complex operations and interactions of its internal components.
Thingness is not an isolated concept. In Knappett and Malafouris (2008), thingness and objecthood exist as a pair of attributes of an entity (including both human and non-human), constantly transforming into each other. When a part of an entity’s thingness is named or labeled, making it visible and perceivable, it transforms into objecthood. Knappett and Malafouris call this process objectification. From this perspective, I believe that the public’s, and even many researchers’, conceptualization of AI is a process of objectification.
Therefore, the phenomenon pointed out by Bucher (2025) is that AI as an object obscures the many underlying characteristics of AI as a thing, as revealed by Suchman (2023). I plan to dedicate a separate note to further explore these ideas.
Tags: ai ant new_matarialism References: Reading List of AI and Education
Knappett, C., & Malafouris, L. (Eds.). (2008). Material agency: towards a non-anthropocentric approach. Springer Science & Business Media.