{"count":7,"registry":"C.R.E.E.D. Remediation Registry","remediations":[{"category":"AI Model Constraints","description":"Local language models reason from their training corpus; they cannot look up events that post-date that corpus or query a live fact store. This is a property of how the models are run, not a configuration bug.","limitation":"LLM analysis is based on training data (cutoff 2024) \u2014 models do not have internet access or real-time fact databases.","owner":"D.A.R.W.I.N. \u2014 Research & Development","remediation_plan":"Facts are taken from the live article text, not the model's memory: summaries are extractive (drawn from the published body), and the LLM is used for framing, tone and bias interpretation rather than as a source of factual claims. A retrieval-augmented fact-check path (grounding on a live evidence store) is on the research roadmap.","slug":"wtf-llm-training-cutoff","source_product":"What The Fact","status":"accepted","status_label":"Accepted Trade-off","target_date":null,"title":"LLM analysis is bounded by the model's training cutoff","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-llm-training-cutoff"},{"category":"Bias & Fairness","description":"A single model's reading of loaded language is one signal among several. Reasonable analysts \u2014 and reasonable models \u2014 can disagree on where a given article sits on the spectrum.","limitation":"Bias scoring reflects the LLM's interpretation of language patterns; it may not match human expert consensus in all cases.","owner":"B.I.A.S. \u2014 News & Media","remediation_plan":"Bias is computed from multiple independent signals (LLM language analysis, outlet editorial-stance priors, and a separate DistilBERT NLI model) rather than one verdict, and the rationale is shown so readers can judge it themselves. Ongoing calibration measures the scorer against outlet editorial baselines.","slug":"wtf-bias-vs-human-consensus","source_product":"What The Fact","status":"in_progress","status_label":"In Progress","target_date":null,"title":"Bias scores may not match human expert consensus","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-bias-vs-human-consensus"},{"category":"Data Coverage","description":"When the extractor cannot confidently locate a story, it attributes it to the publishing outlet's province rather than guessing a precise location it is not sure of.","limitation":"Geolocation uses source province as fallback when AI confidence is low (<0.5).","owner":"L.O.C.A.L. \u2014 News & Media","remediation_plan":"Province-level fallbacks are labelled transparently rather than presented as precise geolocation. The L.O.C.A.L. agent's geo-extraction is being improved to raise the share of articles that clear the confidence threshold.","slug":"wtf-geolocation-province-fallback","source_product":"What The Fact","status":"in_progress","status_label":"In Progress","target_date":null,"title":"Geolocation falls back to source province at low confidence","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-geolocation-province-fallback"},{"category":"Throughput & Coverage","description":"Full LLM analysis is compute-bound. To keep the feed current, each ingest cycle prioritises the highest-signal articles for deep analysis and applies lighter heuristic + editorial scoring to the rest.","limitation":"Not all articles receive full AI analysis \u2014 only priority articles (top 20/cycle) get LLM treatment; all others receive heuristic + source-editorial bias scoring.","owner":"F.A.C.T. \u2014 News & Media","remediation_plan":"Analysis throughput is being scaled across the agent fleet and a backlog drains lower-priority articles toward full coverage over time. Articles that have not yet had full LLM analysis are scored by transparent heuristic + outlet-editorial methods in the interim.","slug":"wtf-priority-article-coverage","source_product":"What The Fact","status":"in_progress","status_label":"In Progress","target_date":null,"title":"Not every article receives full LLM analysis","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-priority-article-coverage"},{"category":"Bias & Fairness","description":"An outlet's bias profile is a moving aggregate of its analysed coverage. A newly added outlet has little history, so it starts from a documented editorial-stance prior and converges as articles accumulate.","limitation":"Outlet bias profiles are rolling averages; new outlets are seeded from editorial stance data and refined over time.","owner":"B.I.A.S. \u2014 News & Media","remediation_plan":"This is the intended design: profiles update as evidence accumulates rather than being fixed labels. Seed priors are sourced from published editorial-stance research and are superseded by measured behaviour as the sample grows.","slug":"wtf-outlet-bias-rolling-average","source_product":"What The Fact","status":"accepted","status_label":"Accepted Trade-off","target_date":null,"title":"Outlet bias profiles are rolling averages, seeded for new outlets","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-outlet-bias-rolling-average"},{"category":"Data Coverage","description":"Where a publisher gates the article body behind a paywall, the full text is not lawfully or technically available to ingest, so analysis is limited to the headline and the publicly-served summary metadata.","limitation":"Paywalled articles (Postmedia, Torstar) have limited body text \u2014 AI analysis relies on title + Open Graph description only.","owner":"F.A.C.T. \u2014 News & Media","remediation_plan":"Paywalled items are analysed only from the publicly-available title and Open Graph description, and this reduced basis is disclosed rather than hidden. We do not bypass paywalls to obtain body text.","slug":"wtf-paywalled-body-text","source_product":"What The Fact","status":"accepted","status_label":"Accepted Trade-off","target_date":null,"title":"Paywalled articles are analysed from title + summary only","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-paywalled-body-text"},{"category":"Bias & Fairness","description":"The supplementary DistilBERT natural-language-inference model was trained primarily on US data, so its read of Canadian political framing can be imperfect.","limitation":"DistilBERT Signal 3 (US-trained NLI model) may not map cleanly to Canadian political framing; treated as a second opinion, not ground truth.","owner":"V.E.R.I.F.Y. \u2014 News & Media","remediation_plan":"Its output is weighted as one second-opinion signal, never as ground truth, and is cross-checked against the LLM and editorial signals. A Canadian-framing fine-tune is planned to improve its domain fit.","slug":"wtf-distilbert-canadian-framing","source_product":"What The Fact","status":"in_progress","status_label":"In Progress","target_date":null,"title":"The DistilBERT signal is US-trained and used as a second opinion","updated_at":"2026-05-31","url":"https://creed.kytranempowerment.com/remediation#wtf-distilbert-canadian-framing"}],"status_counts":{"accepted":3,"in_progress":4},"status_meta":{"accepted":{"color":"#8b5cf6","label":"Accepted Trade-off"},"in_progress":{"color":"#22d3ee","label":"In Progress"},"planned":{"color":"#f59e0b","label":"Planned"},"resolved":{"color":"#10b981","label":"Resolved"}}}