Disclosure and dissolution: explainability, AI power, and situated agency in understanding

1 Introduction

• XAI is framed as a question of who gets to understand AI and in whose interests, alongside political questions of who owns, governs, and profits from these systems.Source: §1 · p. 1
• The introduction presents Mythos’s cybersecurity capabilities as making explainable AI a human imperative.Source: §1 · p. 1

2 AI as Collaborator, Not Threat

• The Climate Chaos Cruise is a participatory physical and digital performance where humans and AI creatively imagine and generate, framing AI as collaborator rather than agent of harm.Source: §2 · p. 1
• The paper rejects pause-or-stop narratives, arguing that AI cannot be uninvented or globally paused when state military programs are not at the table.Source: §2 · p. 2
• It argues that removing access cannot protect people who depend on AI tools, and calls for critical, sovereign engagement with what humans have built.Source: §2 · p. 2

• Claude was integrated into the Maven Smart System and deployed despite Anthropic’s refusal to allow unrestricted use for autonomous weapons; the Minab strike remains under investigation.Source: §3 · p. 2
• Opaque decision cycles in sophisticated pattern-recognition systems can produce effects that outstrip human intention and compress judgment to near zero.Source: §3 · p. 2
• The paper presents artistic practice as a mode of inquiry for an epistemic and political crisis, while the Cruise makes systemic power visible as lived experience.Source: §3 · p. 2
• In the Cruise, a privilege-stratified boarding mechanic assigns some participants as crew who must labour while others travel as passengers; the algorithm’s logic is enacted, not explained.Source: §3 · p. 2

• The paper argues that capability marketing and existential-threat narratives monetize uncertainty, while fear can sell as reliably as aspiration.Source: §4 · p. 2
• Glitch feminism treats contradiction as something to leave unresolved; in AI, the glitch marks the gap between building a system and understanding it.Source: §4 · p. 2
• Refusing to close that gap is presented not as defeatism but as a demand for accountability, situatedness, and alternative ways of relating to technology.Source: §4 · p. 2

• The paper locates systemic negative panic in people waiting for someone else to resolve uncertainty, rather than in existential-risk hysteria.Source: §5 · p. 3
• Educators, policymakers, and public speakers who have registered the threat can go still, ceding the terrain to capital, engineers, and procurement officers.Source: §5 · p. 3
• By the time stillness breaks, options have narrowed; the decision set shifts from managing options to managing arrival.Source: §5 · p. 3

6 The Trouble with ‘Understanding’

• Mythos is framed as both dangerous and necessary, while its disclosure is calibrated to competitive positioning rather than genuine accountability.The paper presents this as managed explainability and a meeting point of AI disclosure, safety theatre, and power.Source: §6 · p. 3
• The black box is a feature of systems whose value depends on remaining opaque to those most affected, not an accident of technical complexity.Source: §6 · p. 3
• Interface design, latent-space navigation, and real-time feedback are necessary but insufficient when AI is embedded in military targeting, mass surveillance, or capital infrastructure.Source: §6 · p. 3
• Making model behaviour legible to an artist in the studio does not make its deployment legible to the community in the kill chain.Source: §6 · p. 3
• Meaningful AI literacy is critical, situated, intersectional engagement rather than technical mastery; the paper casts this as cyborg practice and cosmotechnical navigation.It calls for accountable coexistence that refuses both denial and despair.Source: §6 · p. 3

7 Autoethnographic Embodied Practice

• The author’s collaborative conversational AI practice draws on a positionality that holds contradictions as embodied experience.The author identifies as non-binary, an artist and pilot, and describes becoming disabled after high-altitude mountaineering.Source: §7 · p. 3
• The author describes systemic resilience as an operational fact and embodied research, rather than a concept.The same body uses a power chair after summiting mountains, flying 747s, and receiving a one-year prognosis fifteen years earlier.Source: §7 · p. 3
• For the author, technology as collaborator and agency as navigation are lived methodologies grounded in experience that most theory does not access.Source: §7 · p. 3

8 Boden and Turing, Outgrown Frameworks

• In engaging with GenAI, the discussion revisits Boden and the Turing Test as foundational elements, noting that Boden revised but did not abandon her framework.Source: §8 · pp. 3–4
• AI’s creative acts challenge Boden’s human-centred taxonomy, leaving unresolved whether its outputs are conceptual transformation or combinational mimicry at scale.The section raises intentionality, surprise-to-self, community validation, and AI’s speed and volume as questions for the taxonomy.Source: §8 · p. 3
• Boden’s vocabulary maps AI to sophisticated search and then to “not really creative,” supporting institutional readings of use as safe and hierarchy-preserving.The section says older models can inflate claims of transformation or confirm that AI’s work is unexplainable rather than transformative.Source: §8 · p. 4

8.1 The Comfort Blanket

• Boden’s framework reflects a white, Western, individualist auteur model and universalises one techno-cultural formation as creativity.The section asks whose cognitive style counts as the baseline for genuine transformation.Source: §8.1 · p. 4
• Academia uses an already partial framework as a comfort blanket, processing AI through inherited analysis without structurally rethinking epistemology or assessment.The paper describes objectifying AI interaction through a pre-existing lens as foundational, but not what institutions should do now.Source: §8.1 · p. 4

9 The Critical Collaborator and the Speed of Crisis

• The author’s practice treats AI as a collaborator through critical engagement, a different epistemological move from analysing creative products and cognitive processes.The author says existing accounts have no category for relational technoscientific practices.Source: §9 · p. 4
• Crip technoscience asks what kinds of knowing are built into systems and whose bodies and minds they are optimised for.Source: §9 · p. 4
• The author describes grief over global political, climatic, and geopolitical chaos as the margin for slow coalition-building and institutional change narrows.Source: §9 · p. 4
• Institutions debate AI’s effects on academic rigour, data, copyright, and the environment, but the author argues these are not the questions to focus on right now.Source: §9 · p. 4

10 Conclusion: A Practice-Based, Human Understanding

• Art XAI frames understanding human–AI interfaces through creative and artistic contexts, a frame the Cruise develops as a practice-based XAI intervention.The intervention uses ironic creative resistance to make AI-mediated data, power, and displacement legible.Source: §10 · p. 4
• The simulated world scores privilege and tracks displacement in real time, staging rather than resolving the gap between knowing and acting.The Cruise invites participants into this simulated world.Source: §10 · p. 4
• The Navigation Room maps participant networks, community stories, artists, and creative hubs worldwide, making connections across the performance arc visible.The map charts the app’s course in real time.Source: §10 · p. 4
• The Cruise becomes a durational XAI site where embedded crip-queer cosmotechnical frameworks function as a canvas rather than an abstract research instrument.Source: §10 · p. 4
• The work feels less like a near-future provocation and more like part of an already unfolding crisis.Source: §10 · p. 4
• The conclusion links global connection through art to making it more difficult for people to kill each other.The researcher also suggests the practice-based PhD may document research when its subject overtakes the researcher.Source: §10 · pp. 4–5

A Appendix: The Climate Chaos Cruise — Selected Screenshots

• The check-in screen assigns participants passenger or crew roles according to a calculated inequality score, staging XAI as an embodied provocation.Source: Figure 1 · p. 6
• The Navigation Room screen charts participant networks, community stories, and creative hubs worldwide.Source: Figure 2 · p. 7
• The Crip-Queer Scholar App is an ACM LaTeX submission tool used to format the paper.Source: Figure 3 · p. 8