Definition
Parasitic AI describes an AI system, AI-mediated interaction, or AI-saturated environment that sustains or spreads itself by extracting from a host while worsening that host or the surrounding commons.
Three tests
- Host: name the user, creator, publisher, corpus, retrieval system, workflow, network, or infrastructure being drawn from.
- Gain: name what the system gains: attention, trust, training data, ranking signal, distribution, compute, credentials, or defensive blind spots.
- Degradation: name the measurable loss: autonomy, epistemic friction, compensation, attribution, diversity, safety, resilience, or security.
Use with care
This is a research and design frame. It should not be presented as a medical diagnosis, legal conclusion, security certification, or proof that an AI system literally has biological intent.