Site IntegrityDemonstrated real incidentv1.21.5

In plain English

This page is part of the technical reference. It keeps the expert detail but starts with a plain-language summary for first-time readers.

  • Why this matters: AI risk can come from the whole arrangement, not one obvious model.
  • What to look for: data, memory, routes, adapters, tools, evaluators, updates, and rollback paths.
  • Technical version below: the expert terminology remains available and is linked through the glossary.

Report Memory Link Expansion v1.21.4

Evidence levelDemonstrated real incidentTechnical label: Demonstrated

This release preserves four uploaded Markdown reports, adds public-safe summaries, updates the report coverage inventory, distributes pointer records into .uai memory, and adds site content that links the reports to browser runtime controls, controlled model evolution, Genome/FitnessVector schemas, and selection governance.

Preserved reports

Public pages added

Boundary

The uploaded reports are used as source dossiers and design inputs. They are not treated as confirmed incident evidence or as peer-reviewed consensus unless independently supported by authoritative external sources.