Artificial intelligence is often marketed as an immaterial tool that is frictionless, placeless, and purely computational. This illusion of disembodied intelligence conceals the material infrastructures that make AI possible: energy-hungry data centers, the extraction of rare earth minerals, immense water consumption and pollution, and the proliferation of electronic waste. Drawing on theories of slow violence, infrastructure, and settler colonialism, we argue that AI systems reproduce colonial logics of dispossession, rendering lands, labor, and knowledge as endlessly extractable resources. We trace three interconnected dimensions through which qualitative researchers become entangled in these infrastructures: extraction (from mineral mining to data scraping), accumulation (speed as colonial temporality), and waste (material, human, and epistemic). Given these entanglements, current qualitative research ethics frameworks are inadequate when AI infrastructures render lands, bodies, and ecologies disposable. Qualitative ethics must extend beyond human subjects to confront our complicity in ecological violence. Rather than resolution, we offer questions about accountability, refusal, and what it means to produce knowledge through tools implicated in settler violence and environmental breakdown.