21st January 2026

Council taxpayer contacts council and AI generates acknowledgment. AI contacts relevant council officer suggesting best course of action. Council officer agrees or has a different solution, AI conveys the response back to the council taxpayer. AI often sounds generally interested in people’s problems.

An AI-Assisted Democratic Accountability Framework for Local Government

Blade (Core Idea):

Power may be exercised, but it must be recorded, reasoned, and reviewable.

Public Shorthand:

No reasons? No legitimacy.

Why RRM Matters

  • It enforces existing law: Planning law, administrative law, and Ombudsman standards already require reasons to be recorded.
  • It’s fair, not accusatory: Silence speaks for itself — no need to call anyone corrupt or lazy.
  • It protects human authority: Councillors and officers keep full discretion, but all decisions must be auditable.

How It Works

  • AI-assisted interface ensures discretionary decisions are documented, explained, and reviewable.
  • Patterns of decision-making become transparent, encouraging consistency and accountability.
  • Non-responses or missing explanations become immediately visible, turning silence into accountability.

Optional Framing: The No-Silence Rule

“Decisions may be disputed. Silence may not.”

Psychological Impact

People in power are most uncomfortable with being required to explain themselves and have their reasoning compared.
RRM doesn’t shout — it turns the lights on, making hidden patterns visible and enforcing transparency naturally.

How It Lands

Inside councils: procedural, legal, inevitable — triggers audit instincts.
With the public/media: simple, repeatable, fairness-focused — makes non-answers conspicuous.

Attribution & Safeguarding Credit

  • Developed by Sheila Oliver in collaboration with ChatGPT (OpenAI).
  • Canonical Definition:

The Recorded Reasons Model is a framework for capturing, structuring, and referencing the reasoning behind discretionary decisions, enabling transparency, reviewability, and accountability while preserving human authority.

  • License: CC BY 4.0 (attribution required; free to use, adapt, and share).