Findings by audience

The same hallucination findings, viewed by who is affected. Each case study aggregates findings across the regulations relevant to that audience in that jurisdiction.

Last updated 14 Jun 2026 · Full ledger: Hallucination Register · Companion view: Findings by regulator

RegLegBrief organises every finding into three audience views so the same underlying research surfaces in the form each reader actually uses. Practitioners are individual professionals — lawyers, compliance officers, risk leads, accountants, internal auditors, financial advisers, company secretaries, public auditors — whose work involves reading and applying regulation. Sectors × Departments are the firm-side cross-cuts: a banking risk team reads the same regulation differently from an asset-manager compliance team or a payment-institution operations team, so RegLegBrief publishes cells per (sub-sector, department, regulation) combination. AI Labs are the frontier-model providers (Anthropic, OpenAI, Google, others) whose models are the subjects under test — each regulation gets its own AI Labs whitepaper documenting the failure modes the model exhibited.

The underlying findings are identical across audience views; the framing, interpretation, and "what this means for you" sections are tailored to the reader's role. For the methodology that produces every finding and the rules that govern publication, see /methodology/.