Section 377 of 440
Complete canonical tutorial. This reader section contains the same teaching body as PWR-195 · AI-assisted creativity. Open the Power dossier.
PWR-195 · SELF GUIDED full tutorial
Use AI to vary a human-authored concept while tracking provenance, similarity and ownership
Begin the library-bookmark task with three human sketches, then use the supplied prompt to request six constrained text variations. Track provenance, remove near-duplicates, select through the rubric and produce a fresh bookmark idea later without AI. A traceable selected design is the result; output volume alone is not creativity.
1 · Permission and limits
Know exactly what you may do
2 · Get ready
Gather what you need and check the starting conditions
What you need
- Declared AI-assisted creativity fixture: Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations.
- Setup aid for Create human-first options: Exclude requests to imitate living artists, copy protected brands or expose confidential work.
- AI-assisted creativity log: Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance; retain AI-assisted creativity errors, assistance, stop and fallback.
- Creativity evidence ledger: record ownership and meaning outcomes from “Relying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects”. Add individual novelty and collective-diversity outcomes from “Generative AI enhances individual creativity but reduces the collective diversity of novel content”. Use the NIST Generative AI Profile to document prompt, output, selection and provenance for the bookmark.
- Exact bounded-generation prompt: “Create six one-sentence, text-only concepts for a 55×180 mm fictional library bookmark that directs a first-time visitor from the entrance to Returns Desk 1. Vary the navigation mechanism or wording, keep print readable, use no brand, artist, copyrighted character or personal data, and label outputs V1–V6. Do not score them.” Record the model/version and this prompt verbatim.
- Fixed captured-output pack when no approved AI is available: V1 a large arrow reading “Returns → Desk 1”. V2 a dotted arrow trail ending at “Desk 1”. V3 a three-shelf mini-map with Desk 1 circled. V4 stacked arrows saying “entrance, shelves, returns”. V5 the question “Returns”? followed by “Follow shelf numbers to Desk 1”. V6 a floor-line metaphor ending at a desk icon. These are inputs to cluster and score, not endorsed answers.
- Rubric and delayed prompt: score clarity, usefulness and distinctiveness from 0=misses the brief, 1=partly meets it, 2=clearly meets it; total 0–6 and record origin only after scoring. After 48 hours without AI, design three concepts for a wallet-size sign directing a visitor from an elevator to an accessible restroom using only words and simple shapes.
Before you start
- Use a harmless fictional brief and material you own or may reuse.
- Exclude requests to imitate living artists, copy protected brands or expose confidential work.
- Start check for AI-assisted creativity: The brief can reject a pretty but unusable idea.
- Top-of-sheet stop for AI-assisted creativity: Stop if provenance, licence or source status cannot be determined.
3 · The method
Follow these steps in order
- Define the design need
Write user, function, constraints and a three-part rubric: clarity, usefulness and distinctiveness.
Why: The brief can reject a pretty but unusable idea.
Check: The brief can reject a pretty but unusable idea.
- Create human-first options
Make three rough concepts before opening the AI and label the parts you authored.
Why: A human contribution baseline and range are preserved.
Check: A human contribution baseline and range are preserved.
- Request bounded variation
Give the brief and ask for six alternatives that vary layout or wording, not copied brands or living artists’ styles.
Why: Outputs address the same need and avoid imitation requests.
Check: Outputs address the same need and avoid imitation requests.
- Record provenance
Save prompt, model/version, output IDs, edits and any source material used.
Why: Another reviewer can trace what came from whom.
Check: Another reviewer can trace what came from whom.
- Cluster for similarity
Group concepts that share the same metaphor, wording or layout; count distinct clusters rather than raw output count.
Why: Six near-duplicates are not called six diverse ideas.
Check: Six near-duplicates are not called six diverse ideas.
- Evaluate blind
Remove origin labels and score every human and AI-assisted concept against the prewritten rubric.
Why: Selection is not driven by novelty theatre or model prestige.
Check: Selection is not driven by novelty theatre or model prestige.
- Combine actively
Choose one human element and one verified useful variation, then redraw and explain each contribution.
Why: The final object contains a deliberate human decision, not passive acceptance.
Check: The final object contains a deliberate human decision, not passive acceptance.
- Test delayed unaided creation
After 48 hours, solve a new signage prompt without AI and compare range, confidence and ownership with baseline.
Why: Device-on productivity is separate from retained creative ability.
Check: Device-on productivity is separate from retained creative ability.
4 · Worked example
See the whole method used once
Scenario
Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations.
Walkthrough
- Write the bookmark brief and sketch a blue arrow, a shelf-map and a question-mark concept before AI use.
- Ask for six text-only variations and save the exact prompt and model version.
- Cluster four outputs as “arrow plus desk” variants, one as a floor map and one as a short question.
- Blind-score all nine concepts; the human shelf-map and generated short question lead on clarity and usefulness.
- Redraw a final shelf-map with the words “Returns? Follow shelf numbers to Desk 1,” recording both contributions.
- Two days later, design a restroom-direction card without AI and compare its three distinct concepts with the original baseline.
Result
The final bookmark is traceable and rubric-selected, while the similarity audit prevents raw output volume being called diversity. It does not prove general creativity enhancement.
5 · Right and wrong
Compare correct or safer execution with the common wrong version
| Moment | Right / safer | Wrong / riskier | Why it matters |
|---|---|---|---|
| Human baseline | Make several concepts before AI exposure. | Accept the first generated concept as the starting point. | Early AI influence can narrow later exploration. |
| Output count | Count distinct idea clusters when completing the bookmark concept audit. | Call six similar outputs “six novel ideas.” | Assisted outputs can increase individual quality while reducing collective diversity. |
| Provenance in bookmark concept audit | Keep prompts, sources and edits when completing the bookmark concept audit. | Present generated content as solely human-authored. | Hidden provenance creates ownership and trust problems. |
| Active choice | Redraw and justify selected elements when completing the bookmark concept audit. | Copy an output without evaluation during the bookmark concept audit. | Passive reliance can reduce ownership and meaning. |
6 · Common mistakes
Spot the error and apply the correction
| Mistake | Fix |
|---|---|
| Accept the first generated concept as the starting point. | Lock three rough human options first. |
| Call six similar outputs “six novel ideas.” | Cluster by metaphor, layout and wording. |
| Present generated content as solely human-authored. | Attach a contribution note within the bookmark concept audit. |
| Copy an output without evaluation. | Use the blinded rubric and name each retained contribution. |
7 · Practice
Turn the steps into a usable skill
First session
- Bookmark concept comparison: Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations.
- Prewritten design need: Define the design need: Write user, function, constraints and a three-part rubric: clarity, usefulness and distinctiveness.
- Provenance and cluster audit: Record provenance, then Cluster for similarity.
- First-output-bias correction: if “Accept the first generated concept as the starting point.” appears, apply “Lock three rough human options first.”
- Restroom-sign transfer: Test delayed unaided creation: After 48 hours, solve a new signage prompt without AI and compare range, confidence and ownership with baseline.
Repeat plan
Use one harmless design brief every two weeks. Alternate AI-assisted and unaided sessions, include a 48-hour new-prompt test, and review corpus similarity after four briefs before increasing output count.
Progress when
- The brief can reject a pretty but unusable idea.
- A human contribution baseline and range are preserved.
- Outputs address the same need and avoid imitation requests.
- The final work has complete provenance, at least three distinct concept clusters, blinded rubric improvement and no drop in delayed unaided idea range or ownership.
Do not progress when
- Do not continue while this error remains: Accept the first generated concept as the starting point.
- Pause until this correction works: Cluster by metaphor, layout and wording.
- This AI-assisted creativity stop ends the block: Stop if provenance, licence or source status cannot be determined.
8 · Check the result
Measure what changed
Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance
How: Baseline fixture: Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations. Enter success checks from “Define the design need” and “Cluster for similarity”. If “Accept the first generated concept as the starting point.” occurs, apply its named fix; then score Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance on the unused “Test delayed unaided creation” item.
Good result: The final work has complete provenance, at least three distinct concept clusters, blinded rubric improvement and no drop in delayed unaided idea range or ownership.
This does not prove: Boundary for AI-assisted creativity: “Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance” describes only Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations. It cannot establish “AI unlocks unlimited personal creativity”.
Self-check
- Without the example, demonstrate: The brief can reject a pretty but unusable idea.
- Find the fault in this attempt: “Accept the first generated concept as the starting point.” Apply “Lock three rough human options first.”; what changes?
- What evidence in the completed record shows that this is wrong: “Call six similar outputs “six novel ideas.””?
- AI-assisted creativity stop decision: Stop if provenance, licence or source status cannot be determined.
9 · Stop, adapt or get help
Keep the safety boundary practical
Stop and get help
- Stop if provenance, licence or source status cannot be determined.
- Do not submit private client material to an unapproved system.
- Pause if the learner feels unable to reject outputs or reports falling ownership, meaning or self-efficacy.
Accessibility and adaptations
- Use spoken ideation, collage, tactile sketching or a scribe; score the concept against the same brief.
- Limit visible options to three at a time to reduce overload while preserving human-first generation.
10 · Evidence and limits
Why these instructions are here
- primary research
Registered support for AI-assisted creativity: “Relying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects”. It bears on Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance inside the AI-assisted creativity fixture. It does not validate “AI unlocks unlimited personal creativity”.
Relying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects - primary research
Constraint for AI-assisted creativity, drawn from “Generative AI enhances individual creativity but reduces the collective diversity of novel content”: Assisted stories became more similar to one another, and passive reliance reduced ownership, meaning and self-efficacy.
Generative AI enhances individual creativity but reduces the collective diversity of novel content - official guidance
Generative-AI control for AI-assisted creativity: record version and source checks, then keep human review plus failure behaviour visible. Those controls are exercised in “Combine actively” while measuring Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance.
Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile
Limits
- AI-assisted creativity boundary: interpret “Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance” only for Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations.
- A successful result does not establish “AI unlocks unlimited personal creativity”.
- AI-assisted creativity limiting finding: Assisted stories became more similar to one another, and passive reliance reduced ownership, meaning and self-efficacy.
- No perfect-performance claim for AI-assisted creativity: the evidence register does not make “Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance” universal, consequence-free or flawless in Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations.
- Scope remains AI-assisted creativity: Design a fictional library bookmark that helps readers find the returns desk. The learner sketches three human-first concepts, then asks AI for six text-only variations. Recheck the comparator, support and “Blinded novelty/usefulness, corpus diversity, provenance, human contribution, time and delayed unaided performance” after any configuration change.
Open the complete canonical research register
- Primary empirical supportLimiting / contraryRelying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects
Elena Hayoung Lee; Yidan Yin; Nan Jia; Cheryl J. Wakslak · 2026 · Primary research
- Primary empirical supportLimiting / contraryGenerative AI enhances individual creativity but reduces the collective diversity of novel content
Anil R. Doshi; Oliver P. Hauser · 2024 · Primary research
- Limiting / contraryOfficial boundary contextArtificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile
C. Autio; R. Schwartz; J. Dunietz; S. Jain; M. Stanley; E. Tabassi; P. Hall; National Institute of Standards and Technology · 2024 · Official standard
- Limiting / contraryOfficial boundary contextNIST Privacy Framework: A Tool for Improving Privacy Through Enterprise Risk Management, Version 1.0
National Institute of Standards and Technology · 2020 · Official standard
Read the complete evidence interpretation on the Power dossier.
Tutorial delivery controls
Learn, adapt, troubleshoot and resume
Progress is saved only in this browser on this device.
Step-by-step learner mode
Each activity includes its success check, a nearby accessible alternative and an “I’m stuck” correction path. Alternatives preserve the target where possible; when they change the task, Titan labels them as related rather than equivalent.
Define the design need
Write user, function, constraints and a three-part rubric: clarity, usefulness and distinctiveness.
The brief can reject a pretty but unusable idea.
The brief can reject a pretty but unusable idea.
I’m stuck on this step
Reset: Re-read this authored instruction — “Write user, function, constraints and a three-part rubric: clarity, usefulness and distinctiveness.” — and its success check, then attempt only this step.
Possible snag: The result from “Write user, function, constraints and a three-part rubric: clarity, usefulness and distinctiveness.” does not yet meet this declared check: The brief can reject a pretty but unusable idea.
Correction: Return to the start of “Define the design need”, reduce complexity or pace, and repeat only the part needed to satisfy: “The brief can reject a pretty but unusable idea.”
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Create human-first options
Make three rough concepts before opening the AI and label the parts you authored.
A human contribution baseline and range are preserved.
A human contribution baseline and range are preserved.
I’m stuck on this step
Reset: Re-read this authored instruction — “Make three rough concepts before opening the AI and label the parts you authored.” — and its success check, then attempt only this step.
Possible snag: Accept the first generated concept as the starting point.
Correction: Lock three rough human options first.
Possible snag: Present generated content as solely human-authored.
Correction: Attach a contribution note within the bookmark concept audit.
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Request bounded variation
Give the brief and ask for six alternatives that vary layout or wording, not copied brands or living artists’ styles.
Outputs address the same need and avoid imitation requests.
Outputs address the same need and avoid imitation requests.
I’m stuck on this step
Reset: Re-read this authored instruction — “Give the brief and ask for six alternatives that vary layout or wording, not copied brands or living artists’ styles.” — and its success check, then attempt only this step.
Possible snag: The result from “Give the brief and ask for six alternatives that vary layout or wording, not copied brands or living artists’ styles.” does not yet meet this declared check: Outputs address the same need and avoid imitation requests.
Correction: Return to the start of “Request bounded variation”, reduce complexity or pace, and repeat only the part needed to satisfy: “Outputs address the same need and avoid imitation requests.”
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Record provenance
Save prompt, model/version, output IDs, edits and any source material used.
Another reviewer can trace what came from whom.
Another reviewer can trace what came from whom.
I’m stuck on this step
Reset: Re-read this authored instruction — “Save prompt, model/version, output IDs, edits and any source material used.” — and its success check, then attempt only this step.
Possible snag: The result from “Save prompt, model/version, output IDs, edits and any source material used.” does not yet meet this declared check: Another reviewer can trace what came from whom.
Correction: Return to the start of “Record provenance”, reduce complexity or pace, and repeat only the part needed to satisfy: “Another reviewer can trace what came from whom.”
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Cluster for similarity
Group concepts that share the same metaphor, wording or layout; count distinct clusters rather than raw output count.
Six near-duplicates are not called six diverse ideas.
Six near-duplicates are not called six diverse ideas.
I’m stuck on this step
Reset: Re-read this authored instruction — “Group concepts that share the same metaphor, wording or layout; count distinct clusters rather than raw output count.” — and its success check, then attempt only this step.
Possible snag: Call six similar outputs “six novel ideas.”
Correction: Cluster by metaphor, layout and wording.
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Evaluate blind
Remove origin labels and score every human and AI-assisted concept against the prewritten rubric.
Selection is not driven by novelty theatre or model prestige.
Selection is not driven by novelty theatre or model prestige.
I’m stuck on this step
Reset: Re-read this authored instruction — “Remove origin labels and score every human and AI-assisted concept against the prewritten rubric.” — and its success check, then attempt only this step.
Possible snag: The result from “Remove origin labels and score every human and AI-assisted concept against the prewritten rubric.” does not yet meet this declared check: Selection is not driven by novelty theatre or model prestige.
Correction: Return to the start of “Evaluate blind”, reduce complexity or pace, and repeat only the part needed to satisfy: “Selection is not driven by novelty theatre or model prestige.”
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Combine actively
Choose one human element and one verified useful variation, then redraw and explain each contribution.
The final object contains a deliberate human decision, not passive acceptance.
The final object contains a deliberate human decision, not passive acceptance.
I’m stuck on this step
Reset: Re-read this authored instruction — “Choose one human element and one verified useful variation, then redraw and explain each contribution.” — and its success check, then attempt only this step.
Possible snag: Copy an output without evaluation.
Correction: Use the blinded rubric and name each retained contribution.
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Test delayed unaided creation
After 48 hours, solve a new signage prompt without AI and compare range, confidence and ownership with baseline.
Device-on productivity is separate from retained creative ability.
Device-on productivity is separate from retained creative ability.
I’m stuck on this step
Reset: Re-read this authored instruction — “After 48 hours, solve a new signage prompt without AI and compare range, confidence and ownership with baseline.” — and its success check, then attempt only this step.
Possible snag: The result from “After 48 hours, solve a new signage prompt without AI and compare range, confidence and ownership with baseline.” does not yet meet this declared check: Device-on productivity is separate from retained creative ability.
Correction: Return to the start of “Test delayed unaided creation”, reduce complexity or pace, and repeat only the part needed to satisfy: “Device-on productivity is separate from retained creative ability.”
Stop / get help: Stop if provenance, licence or source status cannot be determined.
Correct versus incorrect execution
These accessible process diagrams are built from the tutorial’s own right/wrong teaching. They are not anatomical illustrations and do not add technique beyond the canonical tutorial.
Make several concepts before AI exposure.
Accept the first generated concept as the starting point.
Count distinct idea clusters when completing the bookmark concept audit.
Call six similar outputs “six novel ideas.”
Keep prompts, sources and edits when completing the bookmark concept audit.
Present generated content as solely human-authored.
Redraw and justify selected elements when completing the bookmark concept audit.
Copy an output without evaluation during the bookmark concept audit.
Method-structure checklist
10 of 10 structural checks present
- Ordered, Power-specific instructions — present
- Every activity has a success check — present
- Materials or supplied records are declared — present
- Measurement or assessment rule is present — present
- Tutorial-specific troubleshooting is present — present
- Stopping or escalation boundary is present — present
- Every activity has an adjacent alternative — present
- Correct-versus-incorrect comparison is present — present
- Evidence context is bound to the Power record — present
- Planning metadata is present — present
The method-readiness band and presence checklist assess tutorial presentation and are separate from evidence quality for the underlying Power. They are automated editorial aids, not human approval.
Manual editorial sign-off: Pending. This tutorial must not display a human-approved state until an identified editor signs the exact content hash.