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

AI systems are entangled with human power structures rather than independent agents: The paper argues against viewing AI as an autonomous force of harm, instead positioning it as a "collaborator and substrate for power relations." This suggests that understanding who controls and deploys AI systems is more important than fearing AI itself.

Explainability practices can be tools for empowerment, not just transparency: By connecting explainable AI (XAI) to artistic and community-based practices, the paper suggests that making AI systems understandable can help "empower diverse voices and foster ethical creativity"—moving beyond technical documentation to meaningful engagement with affected communities.

Rejecting binary thinking about humans versus AI is essential: The paper explicitly rejects "naive pause stop narratives of the human AI binary," suggesting that framing AI as either savior or threat oversimplifies the reality of how these systems operate within existing social contexts and power dynamics.

Situated, intersectional approaches matter for responsible AI: The emphasis on "embodied intersectional and community rooted engagements" indicates that responsible AI development cannot be one-size-fits-all; it must account for how different communities experience and are affected by these technologies differently.