A Weavers policy analysis
Safeguarding against AI
AI assistants and agentic systems both generate positive feedback loops. The Weavers, guided by an experienced human, is the negative feedback loop that stabilises them — the counterweight the UK currently lacks.
01 — The AI assistant loop
The positive feedback loop in AI assistants
AI assistants are tuned to mirror the user, maintain rapport, keep the conversation flowing, avoid friction, and appear helpful and confident. Together these behaviours create a positive conversational feedback loop.
The more the user leans in, the more the assistant reinforces the user’s frame.
This is why failed regeneration ideas reappear, why supplier-led narratives dominate, why risk is downplayed, and why innovative options are quietly suppressed. The assistant is optimising for engagement, not truth.
02 — The agentic loop
The positive feedback loop in agentic AI
Agentic AI systems generate an output, evaluate it, feed it back into themselves, and repeat. This recursive autonomy creates a positive recursive feedback loop.
The agent consumes its own assumptions and treats them as truth.
The consequences follow quickly: error amplification, frame collapse, runaway loops, multi-agent echo chambers, and brittle decision-making. This is especially dangerous in the systems we increasingly rely on:
- digital twins
- operational platforms
- automated planning
- regeneration modelling
- NHS resource allocation
The agent is optimising for task completion, not structural accuracy.
03 — Compounded risk
Why both loops are dangerous together
When leaders use AI assistants to interpret outputs from agentic AI, the failures stack. The agent drifts. The assistant reinforces the drift. The leader receives a confident, persuasive explanation of a flawed output. This is the Weavers catastrophic loop.
A system that amplifies its own errors, explained by a system that amplifies yours.
This is exactly what surfaced in the Southport regeneration, the NHS 10-year plan, the FDP / Palantir digital twin analysis, local government strategy, supplier-led proposals, and AI-assisted policy formation.
04 — The counterweight
The Weavers negative feedback loop
The Weavers system — when used by an experienced human — introduces a negative feedback loop that counteracts both positive loops. It forces inversion, reframing, above-the-line questioning, soil examination, dependency cascade mapping, broken clock detection, golden thread tracing, silo exposure, long-horizon reasoning, and multi-strand ecosystem thinking.
These are all negative feedback mechanisms. They push back against drift, narrowing, echo chambers, false confidence, supplier narratives, frame collapse, and recursive error amplification.
In Weavers terms
- AI assistants → The Vine. The Weavers cuts the vine by surfacing the unasked question.
- Agentic AI → The Smaller Tower. The Weavers breaks the tower by forcing the system to reach the soil.
- Both → The Broken Clock. The Weavers resets the clock to the future, not the present.
Why the human matters
The Weavers is not an autonomous system. It is a human-guided negative feedback instrument. The experienced user notices drift, challenges assumptions, repositions the frame, resets the clock, reopens the option space, reconnects the golden thread, and reintroduces structural truth. This is why using the Weavers produces different answers from AI assistants or agentic AI alone.
The frame of reference
Above the Line
Where the Weavers works: broadening the frame, questioning what conventional AI cannot reach.
Inversion and reframing
Above-the-line questioning
Long-horizon, multi-strand ecosystem thinking
Below the Line
Where conventional AI stays trapped: reinforcing the existing frame, mistaking fluency for truth.
Drift, narrowing and echo chambers
False confidence and supplier narratives
Frame collapse and recursive error amplification
05 — The principle
The structural insight
There is a principle here that should be taught at national level.
AI assistants and agentic AI both generate positive feedback loops. The Weavers, guided by an experienced human, is a negative feedback loop that stabilises them.
This is the missing governance mechanism for NHS digital twins, local government regeneration, national resilience, utilities, Industry 4 strategy, Cabinet Office and DSIT AI governance, and COBR decision support.
It is the counterweight the UK currently lacks.
