Carsten Jan Weichelt / Human-Led AI
Better AI content. Better decisions.
The W.E.I.C.H.E.L.T. Method is an eight-stage approach to creating, questioning and evaluating AI-assisted content without outsourcing your judgment.
Framework v2.0 / In development and evaluation
01 / The problem
A better prompt isn't always a better process.
Generative AI can deliver a convincing draft in seconds. It can even give you a beautifully written argument you never intended to make.
I work with AI in writing and content strategy. What interests me is not simply how quickly something can be produced, but how we decide whether it deserves to be published.
The W.E.I.C.H.E.L.T. Method is my developing framework for making that process more deliberate, transparent and useful.
02 / Three distinct dimensions
Output Quality
Is the result accurate, useful, appropriate for its audience and faithful to the intended meaning?
Process Quality
How was the content developed, challenged, revised and checked?
Decision Authority
Who determined the direction, selected the alternatives and approved the final result?
03 / The workflow
Eight stages. One recurring question.
Where does AI help, and where do humans need to decide?
The German-English acronym comes from my name. The workflow is designed to be adapted to real writing and content tasks, not followed like a sacred recipe.
Wirkungsabsicht
Purpose & IntentDefine the purpose, audience, intended outcome and constraints before generating content.
Key decision: What is this work actually meant to achieve?Eigene Perspektive
Human PerspectiveEstablish the human argument, observation, experience or creative intention behind the work.
Key decision: What do I actually want to say?Input & Informationsbasis
Research & ContextGather relevant sources, background knowledge, references and requirements. Check what can be verified.
Key decision: What information can this work rely on?Creation
Generation & ExplorationUse AI where useful to explore ideas, structures, alternatives and possible drafts.
Key decision: Which possibilities are worth considering?Hinterfragen
Critical EvaluationChallenge suggestions, question assumptions and identify factual or unintended shifts in meaning.
Key decision: What should be accepted, changed or rejected?Editorial Refinement
Editing & MeaningImprove structure, clarity, voice and readability without losing the intended argument or creative direction.
Key decision: Does the revision still say what it should?Leistungsprüfung
Quality AssuranceCheck facts, intent, audience fit, editorial quality and any required specialist approvals.
Key decision: Is the work ready, and who can approve it?Transfer & Training
Learning & TransferReflect on the process, identify useful lessons and develop transferable skills for future tasks.
Key decision: What should we learn from this?04 / An illustrative example
The sentence is good. The argument isn't.
Suppose you're writing an article about AI and creative work. You ask AI for alternative hooks. One sounds excellent. Unfortunately, it makes a claim you never intended to make.
Example workflow / Fictional illustration
The original point
AI expands creative possibilities, but people still need to evaluate what they create.
The suggested hook
"AI is taking creativity away from writers."
The rejection
It's punchy. It may even attract attention. But it contradicts the original argument. So it doesn't make the cut.
The revised direction
"AI gives writers more possibilities. Who decides which ones matter?"
This is a constructed teaching example, not a documented experiment or evidence that the method improves performance. Real case studies are being developed.
05 / Evaluation
What should we check?
A useful content evaluation asks more than whether a text sounds convincing.
Does the result preserve the intended argument or purpose?
Are claims supported by appropriate, verifiable sources?
Does the text reflect the intended human or organizational voice?
Is the result useful for the people it is meant to reach?
Is the structure clear, the language appropriate and the argument coherent?
Who selected the direction, verified the content and authorized publication?
06 / The conceptual foundation
AI contributes. Humans decide.
The Human Decision Map examines two different questions: where AI participates in creative work, and who makes the decisions that shape its meaning and direction.
It provides the conceptual foundation for the W.E.I.C.H.E.L.T. Method.
Authorship isn't a word count. It's a pattern of consequential decisions.
Human Decision Map07 / Scope & limitations
What this method doesn't claim.
The approach draws on established ideas in human oversight, editorial quality assurance, AI literacy and human-AI collaboration.
Its proposed contribution is bringing these perspectives together in a practical content workflow.
It is not a scientific test for determining legal authorship.
It does not guarantee accurate or high-quality AI-generated content.
It does not replace subject-matter experts, editorial review or required approvals.
It is not a certification of AI regulatory compliance.
08 / Explore further
The thinking behind it.
More on AI-assisted writing, human judgment and the future of authorship.
A final thought
AI gives us more possibilities. It doesn't make the decisions for us.
The W.E.I.C.H.E.L.T. Method is an ongoing exploration of what better human-led AI content work could look like.