to star items.

P002


1 paper proposal Propose
Hacking the Know-It-All Machine: Contesting AI Authority, Reclaiming Agency  
Convenors:
Xin Zhan (SOAS University of London)
Jake Zaslav (Concordia University)
Send message to Convenors
Format:
Panel

Short Abstract

This panel invites ethnographic accounts of how people hack, play with and challenge AI. It explores how these practices contest AI’s authority, redraw boundaries between human and machine, and reshape people’s sense of themselves as knowing and acting subjects.

Long Abstract

What happens when a machine seems to know better than we do? Governments rely on AI to determine who receives social benefits. Companies use surveillance software to track employee productivity. Millions of people turn to chatbots for therapy. Yet this adoption is also accompanied by experiments that test and sometimes disrupt AI’s authority, from “adversarial clothing” designed to confuse facial recognition software to guides on how to make a chatbot have a “mental breakdown”. When a technological object becomes a “subject supposed to know” (Lacan, 1977), how do people question its epistemic authority and make room for their own judgement?

This panel invites ethnographic research on how people test and challenge AI’s authority. We welcome accounts of hacking, jailbreaking and modifying models, alongside users role-playing, joking with or simply messing with AI to see what happens. In her study of trolling and hacking, Coleman (2014) describes a form of playful provocation that “defies boundaries but also re-erects them”. This prompts us to ask what experiments with AI unsettle as well as what they reproduce. We also welcome examinations of peer groups collectively testing chatbot claims, professionals questioning or overriding AI recommendations, and communities adapting models to their own languages and purposes.

How do people make AI something they can question and change? How do their actions draw boundaries between human and machine, subject and object, “high-agency” and “NPC”? Attending to the pleasures, frustrations and ambiguities of contesting AI, we explore how these practices reshape people as knowing and acting subjects.

This Panel has 1 pending paper proposal.
Propose paper