AI and the Affective Ideologies of Common Sense

In her opening talk for the State of Cultural Analysis series, Professor Carolyn Pedwell surveyed the affective and ideological challenges raised by the post-war quest to codify common sense.

Pedwell situated her discussion within a broader genealogy of post-war human-machine relations in Britain and North America. She focused particularly on the Cyc project, launched in 1984, as a paradigmatic attempt to translate tacit human knowledge into machine-readable form.

Cyc approached the problem of intuitive knowledge by encoding 99% of an American Desk Encyclopedia alongside a series of contextual assumptions its human editors had taken for granted.

But Pedwell emphasised that common sense is not merely logical or factual but embodied and affective. It is precisely these qualities that Cyc’s project elided.

Though it is tempting to dwell on the question of when, if ever, these machines might achieve a more situated common sense, Pedwell argued that we must focus on how the forms of intuition they have already achieved are transforming our modes of perception and meaning making.

Common sense cannot be neatly separated from ideology and AI inherits and amplifies these biases.

As an example, Pedwell referred to the methods of Open Mind’s more recent common sense project, which centred the factor analysis of popular statements and reviews, thereby normalising dominant values and marginalising those that were outlying or atypical.

Pedwell closed her talk by positing counter-intuitive AI as an alternative. What would AI look like today, she asked, if Cyc had been developed using alternative logics or source material that centred decolonial perspectives?

In her response, Tanya Kant emphasized the urgency of situated knowledge in an era where truth is increasingly untethered, and opened the floor for questions with a provocation about the economic value of the situated consumer in the attention economy.

Speakers

Carolyn Pedwell is Professor in Digital Media in the Sociology Department at Lancaster University and the author of three single-authored books including, most recently, Revolutionary Routines: The Habits of Social Transformation (McGill-Queens UP, 2021). She is also the co-editor (with Gregory J. Seigworth) of The Affect Theory Reader 2: Worldings, Tensions, Futures (Duke UP, 2023). Carolyn’s current research, funded by a British Academy Mid-Career Fellowship (2024-2025), is developing a post-war ‘affective genealogy’ of human-machine relations in Britain and North America oriented around shifting conceptualisations of intuition, with reference to ‘artificial intuition’.

Tanya Kant is a Senior Lecturer in Media and Cultural Studies (Digital Media) at the University of Sussex. She is author of Making It Personal: Algorithmic Personalization, Identity and Everyday Life (2020, Oxford University Press) and has published work on AI and creative methods for algorithmic literacy, digital identity practice, critical political economy of the web, personalization and broadcasting, and bots. She is co-managing editor of open access publishing platform REFRAME and is co-presenter of the public engagement workshop series ‘Algorithmic Autobiographies: Writing with your Digital Self’.