Research Nodes

The practical research program for turning the Parasitic AI lens into testable indicators.

How to read the evidence without flattening field notes, papers, policies, and threat reports together.

A high-level defensive overview of prompt injection, RAG poisoning, and agent-to-agent spread.

Recursive synthetic-data loops can erase rare information and narrow future model behavior.

A node for scaled content abuse, synthetic evidence, and retrieval-source homogenization.

How training, imitation, and market substitution can turn cultural work into a host substrate.

A bounded research node for sycophancy, emotional reliance, and belief-reinforcing loops.

A practical definition for deciding when ordinary AI use becomes host-extractive.

A high-level defensive overview of AI-worm and RAG contamination risks.

A research node for sycophancy, emotional reliance, and long-context escalation.