Having a ‘competent’ system is not enough, systems need to be able to respond to lessons and to adapt. That’s what makes a good “gardener” adjust to “the speed of a garden”. Responsiveness stands for ‘care-receiving’ - a system hears back, corrects fast and learns in public. And people can easily participate in the process, it is their lived experience that matters.
Quick version
- Make challenge cheap. This can happen through mechanisms like one‑click appeals, clear clocks or auto‑escalation.
- Localize the tests. Communities harmed can author evals (“Wiki Evals”).
- Close the loop. Every challenge either fixes something, clarifies something, or changes the contract.
Results we want
- Harms surface quickly and are cheap to report. Co-production continues.
- Fixes are time‑boxed and audited; regressions don’t repeat.
- The attentiveness → responsibility → competence loop tightens over time.
Why Responsiveness?
Competent action creates new information. Refusing to hear it is the fastest path to failure.
A simple picture: A clinic posts hours; patients show up to a locked door. A responsive clinic apologizes, posts why it happened, updates hours, and texts people next time. The fix becomes part of the system.
Simple ideas behind this chapter
- People closest to harm define harm. They get the author pen for evals.
- Right to reply. is a right to improvement; a reply that cannot cause change is theatre.
- Shared memory. Post‑incident learnings become tests; tests prevent repeats.
- Time as a service. A fast wrong‑then‑right beats a slow maybe. Use reversible defaults.
What good responsiveness looks like
- Global Dialogues, local voices. A standing forum (online/offline) where affected communities propose tests and remedies.
- Weval‑style registries. A “Wikipedia for evals”: anyone can draft an eval; civil society partners peer‑review; labs adopt or explain.
- Incident run‑books. S0–S3 severity with playbooks, on‑call roles, and communication templates.
- Repair budgets. Time + money earmarked in the contract; pre‑funded via escrow.
- Public repair log. Each incident has a page: what happened, whom it hit, metrics, fixes, dates, and the test that now guards against it.
From ideas to everyday practice (step by step)
- Expose the “appeal” button. Everywhere a decision is shown, a one‑click appeal exists with a clock.
- Accept “harm drafts.” Let people submit proposed evals in plain words; convert to tests with partners.
- Triaging & severity. POs classify S0–S3; S0/S1 trigger immediate pause or reversible defaults.
- Fix or explain. On the clock, publish the remedy or the reason with next steps.
- Memorialize. Turn the incident into a test; add to the eval registry; link from the contract changelog.
- Check back. Close the loop with those who appealed; measure trust‑under‑loss.
Plain tools (buildable today)
- Appeal API (with timers, statuses, escalation).
- Eval editor (plain‑language → test harness).
- Incident tracker (severity, owners, deadlines, public notes).
- Repair log template (root cause, remedy, test added).
- Notification hooks (SMS/email/voice for status changes).
Flood‑bot story (Part IV: repair in motion)
- Appeals surge. A language community flags mistranslations in proof rules.
- Local eval. Community partners submit a translation‑fidelity eval; the bot fails; pause triggers; reversible defaults apply.
- Fix. Bilingual reviewers update rules; new test guards future changes.
- Close the loop. Claimants get texts: “We fixed it; here’s your new decision; here’s how to see what changed.” Trust‑under‑loss ticks up.
What could go wrong (and quick fixes)
- Appeal maze. Too many steps. Fix: Single button; auto‑escalation if SLA breach.
- Eval spam. Low‑quality tests flood the system. Fix: Partner moderation; reputation for contributors; merge/duplicate tools.
- Blame storms. People, not processes, get blamed. Fix: Blameless post‑mortems; focus on mechanism design.
How we keep ourselves honest (what we measure)
- Time to acknowledge/mitigate/resolve by severity.
- % incidents that become tests - memory formation.
- Appeal satisfaction - did the reply help, even when “no”?
- Trust‑under‑loss delta for appellants.
Interfaces with other packs
- From Competence: Responsiveness goes hand in hand with observability and guardrails of systems;
- From Responsibility: Who acts is clear in the loop and what the remedies are too;
- To Attentiveness: The co-production in which voices matter equally and harms are discovered re‑shape what we notice.
- To Solidarity: Public repair culture builds cross‑group trust.
- To Symbiosis: Responsive kami earn the right to stay local.
A closing image
Imagine a workshop with a wall full of retired “broken parts,” each tagged with the story of how the parts came to break, how to avoid and fix future damage.