{
  "slug": "model-entrenchment",
  "url": "/papers/model-entrenchment/",
  "title": "Model Entrenchment: Why Useful AI Systems Become Difficult to Remove",
  "type": "working-paper",
  "typeLabel": "Working paper",
  "lifecycle": "proposal",
  "epistemicStatus": "conceptual-framework",
  "statusLabel": "Conceptual framework · experiments proposed, none run",
  "programme": "deployment",
  "question": "When an AI system resists removal, where does the resistance live: in the model's internal representations and behaviour, or in the web of dependence around it?",
  "date": "2026-07-01",
  "version": "v1.1",
  "authors": [
    "Latent Minds Institute"
  ],
  "methods": [
    "linear probes (proposed)",
    "activation steering (proposed)",
    "counterfactual environments (proposed)",
    "behavioural evaluations (proposed)",
    "case coding (proposed)"
  ],
  "models": [
    "open-weight models (proposed)"
  ],
  "evidence": "conceptual framework with falsification criteria; every empirical claim is a prediction",
  "code": null,
  "data": null,
  "related": [
    "interpretability-map",
    "gpu-credit-underwriting",
    "gpu-forward-market"
  ],
  "provenance": null,
  "canonicalUrl": "https://latentmindsinstitute.com/papers/model-entrenchment/",
  "apiUrl": "https://latentmindsinstitute.com/api/research-objects/model-entrenchment.json"
}