5.4Generative AI as a creative partner
The essentials in 30 seconds
- Amplifier: AI quickly generates variants, angles and provocations — ideal in divergence.
- Limits: it recombines the existing, can produce the banal and errors to verify.
- Stance: AI proposes, humans dispose — human discernment to sort and decide.
Threaded mini-case — Sara and the idea drought
Fourth episode: before her third workshop, Sara is stuck — always the same leads. She tries an AI tool: first “give me ideas” (polite banalities), then by pushing it (“suggest truly unexpected leads”) before sorting herself. This section tells how AI relaunches her divergence without deciding in her place.
Sequel to cases 5.1-5.3 — Sara boosts her divergence with AI.
1. An ideation amplifier
Generative AI (ChatGPT, Claude, Gemini, image generation…) is a powerful creative partner. It can quickly generate many variants, suggest unexpected angles, act as a tireless brainstorming partner, provoke, help reframe a problem.
Well used, it multiplies the divergent phase and unblocks idea droughts — like Sara's before her third workshop.
2. A tool, not a judge
But AI is a tool serving human creativity, not a replacement for judgment. It generates from what already exists (it recombines, it doesn't truly « understand »): excellent for volume and variants, but deep originality, critical evaluation and sense of context stay human.
It can produce the banal if you don't push it, and errors — always verify, never take its outputs at face value.
3. AI proposes, humans dispose
The right stance: use AI to enrich and accelerate ideation (generate, explore, provoke), then apply human discernment to sort, evaluate, contextualize and decide (convergence).
Keep an eye on the stakes (reliability, bias, intellectual property, confidentiality). Tomorrow's creative isn't replaced by AI: they are augmented by it, provided they steer it with discernment.
🔎 Concrete example — Push AI beyond the banal
Quality depends on your prompts and your work on the output.
Passive: « Give me ideas. » → conventional, average suggestions.
Active: « These ideas are classic, suggest truly unexpected paths », then human sorting and refining.Result: AI enriches the material; you work it and decide.
Reflex: verify and keep your critical thinking.
Section vocabulary
- Generative AI (creativity)
- AI tools (ChatGPT, Claude, Gemini, image generation…) able to stimulate ideation.
Simple example: Sara generates 20 hook variants for her workshop in 5 minutes.
What it is not: an infallible judge deciding in your place.
Reflex: use it to generate, never to decide alone.
- Brainstorming partner
- Using AI to generate variants, angles and provocations, like a tireless ideation partner.
Simple example: “Give 10 opposite angles on this client problem.”
What it is not: asking it for a final decision with no human sorting.
Reflex: ask for volume and strangeness, sort after.
- Recombination
- AI's way of working: it produces from the existing, without true understanding or deep originality.
Simple example: Its first leads look like what already exists.
What it is not: believing it “understands” your context like you do.
Reflex: push it beyond its first banal answers.
- Human discernment
- Judgment, critical evaluation, sense of context that stay the person's role (convergence).
Simple example: Sara discards 15 off-context leads and refines 5.
What it is not: publishing AI outputs as they are.
Reflex: sort with your criteria, not the tool's.
- Augmented creative
- A person whose creativity is multiplied by AI, without human judgment being replaced.
Simple example: Sara: drought unblocked in 1 hour, decision kept in hand.
What it is not: delegating all thinking to the tool until atrophy.
Reflex: cultivate your own creativity in parallel.
In practice — Using generative AI as a creative partner
- Use AI to stimulate ideation: generate variants, suggest angles, brainstorm, provoke, reframe.
- Push it (precise prompts, requests for boldness) to avoid the banal; use it mainly in divergence.
- Then apply human discernment: sort, evaluate, contextualize and decide.
- Always verify (errors, bias) and keep the stakes in mind (reliability, intellectual property, confidentiality).
Ready-to-say script — prompt in divergent mode (3 prompts)
"Prompt 1: give me 15 varied leads on [problem]. Prompt 2: these ideas are classic — suggest 10 truly unexpected, even provocative leads. Prompt 3: for my 5 favorites, list objections and blind spots. Then I sort by our context."
Adapt to your topic: volume, then boldness, then critique — human sorting last.
Copy-paste template — 3-step AI session
TOPIC: _______________
1) VOLUME: “15 varied leads on ___” → keep: ___
2) BOLDNESS: “10 unexpected, even provocative leads” → keep: ___
3) CRITIQUE: “objections and blind spots of my 5 favorites” → note: ___
HUMAN SORT: kept ___ / discarded (why) ___ — CHECK: ___Solo in a hurry
20-min session: 1 volume + 1 boldness + immediate sorting. Better a short decided session than long passive listing.
As a team
AI upstream (material for the workshop), collective sorting in session: the tool feeds, the group decides.
Sensitive data
Never paste confidential or client data: anonymize the topic before any prompt to the tool.
Visual / product
Image or mockup generation as visual provocations, then human selection on context criteria.
✓ Best practices
- Use AI mainly in divergence (generate, explore, provoke).
- Push it beyond its first (often banal) answers.
- Verify systematically: AI is an assistant, not an authority.
✗ Common pitfalls
- Taking AI's output at face value (errors, bias).
- Passively accepting banal ideas without working them.
- Relying on AI to the point of atrophying your own creativity.
Key takeaways
- Generative AI stimulates ideation: variants, angles, brainstorming, provocation, reframing.
- It multiplies divergence, but recombines the existing and can produce the banal or errors.
- Judgment, evaluation and sense of context stay the person's role (convergence).
- Stance: AI proposes, humans dispose; an augmented creative, not replaced.
48h challenge — apply tomorrow
Run an AI session in 3 steps (volume → boldness → critique) on 1 real topic, sort 3 leads yourself, and verify each reused fact. Note what AI brought — and what it missed.
- I can define generative AI, recombination and discernment without hesitating.
- I ran 1 volume-boldness-critique session with human sorting.
- I did the 48h challenge or scheduled its date.
Frequently asked questions
Will AI replace human creativity?
No — it transforms and amplifies it. It's remarkable at producing many variants fast, exploring angles, acting as a brainstorming partner. But it recombines the existing without real understanding (it tends toward the conventional if not pushed), and it lacks judgment, sense of the human and context, critical evaluation, intention. It can err and reproduce biases. The human role stays central, especially on convergence (sort, decide). Tomorrow's creative isn't replaced, they are augmented: AI proposes, humans dispose.
How to use AI without banal ideas or loss of critical thinking?
Use it to multiply, not replace your thinking (generate many paths to sort). Push it beyond the conventional (precise, bold requests, iterations). Keep systematic critical thinking: verify, question relevance to your context, evaluate with your criteria. Bring your human value (context, intention, meaning). Don't rely on it to the point of atrophy: keep cultivating your own creativity. The danger isn't using AI, but using it passively. Stay active, demanding and critical.
Sara asks the AI “give me ideas” and wants to publish the answers as they are. What is the right reflex?