Reading room
This page summarizes the reviewed improvement handoff research without treating raw intake as final public truth. It groups sources by use: naming the phenomenon, measuring mechanisms, managing legal and policy risk, and defending AI-enabled infrastructure.
Source map
Read every source at the right strength.
Parasitic AI and persona-parasitology essays
Useful for naming patterns, seed prompts, and anecdotal case clusters. Treat as hypothesis-generating.
Sycophancy, sensitive-conversation, and model-release notes
Useful for deployment failures, mitigation direction, and provider-side safety vocabulary.
Model collapse, retrieval collapse, prompt injection, long-context safety, and mental-health chatbot studies
Useful for mechanisms, benchmarks, and reproducible evaluation design.
NIST AI RMF profile, EU AI Act transparency rules, copyright reports and litigation
Useful for governance posture, source transparency, and market-structure questions.
AI misuse, agentic security, RAG poisoning, and non-human identity reports
Useful for defensive indicators, but often telemetry-dependent and incomplete.
Research nodes
Continue into the research archive for pathway-specific notes on persona capture, creative extraction, retrieval pollution, synthetic contamination, and agentic propagation.