Human in the loop

Garden · not yet reviewed Established: widely used

Designing AI work so a person reviews or approves at the points that matter, especially before anything is sent, published, paid or deleted.

Also called: HITL · approval steps · human review · confirmation before acting · sign-off

Connections

Understand first

Ideas this one builds on.

• Steering (garden)

Related

Often used together or easily confused.

• Automation (garden)
• Privacy and data (garden)
• Agents (garden)
• Allocation economy (garden)
• Centaurs and cyborgs (garden)
• Jagged frontier (garden)
• Productive friction (garden)
• Prompt injection (garden)

Skills that use it

• practice-coach (garden)
• /expert-council (garden)

Blueprints that practise it

• Write a handover brief before you give an agent a real task (garden)
• Ask AI to play devil's advocate on a team decision (garden)
• Teach your AI assistant your team's jargon, one confirmed term at a time (garden)
• Turn messy notes into a status draft you can check (garden)

Resources that teach it

• How AI assistance impacts the formation of coding skills (garden)
• AI Fluency: Framework & Foundations (garden)
• How to Improve Interview Feedback Consistency with a Custom GPT (garden)
• Claude Cowork for PMs: My Self-Improving Productivity System (How I AI, with Daniel Blum) (garden)
• The Knowledge Economy Is Over. Welcome to the Allocation Economy. (garden)
• Compound Engineering (garden)
• The Eight Levels of AI Adoption (garden)
• We Gave Every Employee an AI Agent. Here's What Happened. (garden)
• An Executive's Guide to Implementing AI (garden)
• Securing an Always-on AI Employee (garden)
• The Impact of Generative AI on Critical Thinking (garden)
• A Conversation with Dr. Vivienne Ming, author of Robot-Proof (garden)
• In Conversation... with Vivienne Ming (garden)
• Robot-Proof: When Machines Have All the Answers, Build Better People (garden)
• Assigning AI: Seven Approaches for Students, with Prompts (garden)
• Centaurs and Cyborgs on the Jagged Frontier (garden)
• Co-Intelligence: Living and Working with AI (garden)
• The Cybernetic Teammate (garden)
• Management as AI superpower (garden)
• An opinionated guide to which AI to use to do stuff (Summer 2026 Edition) (garden)
• On Working with Wizards (garden)
• Use My /Human-Review Skill to Edit HTML and Markdown Files Visually (garden)
• Evaluating AI for high-stakes work (garden)
• Grok Bot use cases (garden)

Competency

• Steer
• Use responsibly

Find human in the loop on the commons map ↗

You may know it from

• finance sign-off authority Spending above a threshold needs a named person's approval, however routine the request.
• aviation a pilot monitoring the autopilot The system flies; the pilot watches, and takes over when something looks wrong.

In products

The same idea, under different names in different tools.

Provider	What they call it
OpenAI	ChatGPT agent confirmations before actions
Anthropic	Claude permission prompts
Peter Yang	/human-review skill: edit and comment on AI output in the browser, then send all feedback at once

In plain words

The more AI can do on its own, the more it matters where a person steps in. Human in the loop means deciding, on purpose, which steps a person must review or approve. The aim is a few well-placed approvals.

A useful way to choose: look for steps that are hard to undo or that affect other people. Sending an email to a client, publishing a post, paying an invoice, deleting files and making decisions about someone’s job, money or health all deserve a human check. Drafting, sorting and summarising usually don’t need approval each time, though they need occasional spot checks.

Approval only works if the person reads what they approve. Rubber-stamping fifty AI drafts a day without reading them is a loop in name only. Keep approval steps few enough that they stay meaningful.

Example

An agent prepares weekly reminder emails to families about overdue library books. It drafts and queues them, but a staff member reviews the list each Thursday before sending, because one wrong email to the wrong family would damage trust.

How it connects

• A structured form of steering.
• The safety design for agents and automation.
• Required wherever privacy and data or people’s welfare is at stake.

Try it

Pick a task you’d like AI to do for you regularly. Write down each step, then circle the one you’d never want done without you seeing it first. That’s your approval point.

---

Read this card as plain Markdown: /concepts/con-human-in-the-loop.md. Know a parallel from your own field, or a product that implements this? Add it.