Zum Inhalt springen

English:Creative AI, authorship and responsible collaboration - AI Year 5

Aus MOOCsWiki Staging
aiMOOC-Siegel aiMOOC

Creative AI, authorship and responsible collaboration - AI Year 5

QR-Code



Introduction

Books from different language communities on display in a library
Plan an original reading-corner poster with a clear contribution and source record.

Good creative collaboration leaves a clear trail: who drafted, what helped, what changed and what still needs checking. Revise a short poster message, trace six fictional review records and give the teacher meaningful control. No live pupil AI account or public upload is required.


Create a brief and a contribution trail


Start with an audience and a creative purpose

A fictional class is designing a small reading-corner poster. Its audience is other learners choosing a book. The goal is an inviting, readable message, not the most complicated technology. Decide what the poster should help someone do: choose a book and begin a conversation. A human can write alternatives directly; an approved AI tool might sometimes suggest options, but it must earn its place in the workflow.

This course uses original prepared text and fictional review records. It requires no pupil AI account and includes no live generated-image task. A teacher-prepared suggestion simulates one possible kind of assistance so that everyone can practise editing and documenting decisions without sending personal work to an external service.

CodeAI: How Chatbots and Large Language Models Work. A teacher can select a short explanation of generated suggestions and ask why a plausible sentence still needs human judgement. The wider unit is for older grades. Our Year 5 studio remains a supervised paper/text activity; the film does not authorise signing up for its wider live activities.


Keep a contribution record rather than hiding the process

Here is an original fictional version history. A learner-role draft says: Shelves hold journeys we have not taken yet. A separate teacher-prepared assistance card says: Step into tomorrow at every shelf. That second line is deliberately written for this lesson; it is not an actual output from a named AI product. Do not invent a transcript or claim a tool produced it.

The editor-role final choice is: Choose one book. Share one question. It gives the audience two concrete actions and avoids a sweeping claim about the library. These are editorial choices, not a scientific finding that one line is always better. Different audiences may prefer another defensible original wording. Explain the reason for your choice and keep the earlier drafts.

Version Contribution Human decision
V1 Original learner-role line about journeys Retain as a draft; assess clarity for the audience
Assistance card Teacher-prepared suggestion about tomorrow Record it as simulated assistance, not an actual tool output
V2 Original editor-role line with two actions Explain the revision and keep the contribution trail


Build a balanced creative brief

Name the audience, purpose, tone, format and constraints. For this poster, use a short message, readable contrast, room for a credited image and a text alternative. Do not ask for a famous artist's exact work to be copied or claim the poster has that person's endorsement. A real tool's attractive output would still need review for suitability, provenance and school approval.

Compare two workflows: write two original human alternatives, or evaluate two teacher-prepared assistance cards. Use the same audience criteria. Do not declare the assisted route faster or better without measuring and reporting an appropriate comparison. A larger list of suggestions can increase the amount of checking rather than improve the final decision.

A helpful diagnostic asks: who drafted, who suggested, who selected and who verified? These roles may overlap, but they are different actions. An editor who checks a suggestion contributes judgement; copying a suggestion unchanged does not justify claiming to have independently written every word. Honest process notes should identify actual assistance without inventing legal ownership conclusions.



Trace sources permissions and review records


Inspect the source before choosing a visual

The real library photograph above is credited on its Commons file page to Roc0ast3r under CC0. It supports the reading-corner theme. It is not a photograph of our fictional class, proof of who wrote the poster or a licence for every book cover in any other photograph. Source identity, licence and intended use must be checked for the particular material.

For an open-licensed image, retain the original source, creator, exact licence/version and a note of changes. A statement such as found online is not a source record. Credit is not the same as permission: naming a creator cannot by itself make an otherwise unauthorised use permitted. A licence badge copied into a draft is not proof that the person displaying it could grant all needed rights.

Creative Commons BY-SA 4.0 permits reuse subject to its terms, including appropriate credit, a licence link and an indication of changes. Adaptations have ShareAlike conditions. The licence does not settle every possible privacy, publicity or other right. Ask responsible staff about an uncertain case instead of treating a classroom checklist as a legal determination.


A review queue with four explicit flags

Our small program organises six fictional records for a teacher. It cannot inspect copyright ownership, read an actual licence, judge creativity or publish a poster. Staff would need evidence behind each flag. A false flag entry can make a bad record look ready; that is a limitation, not a permission shortcut.

The four fields are private_data, school_ok, rights_checked and credit_complete. Each is a true/false entry in the supplied worksheet. Private_data marks unnecessary identifiable information that this task must stop using. School_ok records review of this proposed school workflow. Rights_checked records that responsible staff have assessed the actual intended material/use. Credit_complete records the required source/contribution information. These simplified flags are not a complete real compliance checklist.

Record Scenario private_data school_ok rights_checked credit_complete
A1 Original caption false true true true
A2 Open-licensed image record false true true true
A3 Unverified web image false true false true
A4 Unnecessary private information true true true true
A5 Incomplete credit record false true true false
A6 Unapproved service proposal false false true true

The rules run in order: stop private information first; otherwise hold an unapproved school workflow; otherwise hold unclear rights; otherwise hold incomplete credit. Only a record that passes these routing checks reaches READY_FOR_TEACHER_REVIEW. That name is deliberate: a person still checks the actual evidence, suitability and final use. There is no automatic publish or legal-clearance output.


Trace each of the six cases

A1 is an original fictional caption with all review records supplied, so it reaches teacher review. A2 is a fictional open-licensed image record whose intended use and credit have already been checked by the responsible role; it also reaches review. These outcomes do not mean all original work or all CC material is automatically suitable for every purpose.

A3 has a source-looking credit but no verified rights assessment, so it waits for that assessment. A4 includes unnecessary private information, so the stop takes priority even when the other flags say true. A5 has a missing credit element, so the record waits for a complete source/contribution note. A6 proposes a service not approved for this school activity, so it waits for staff rather than inviting pupils to try it anyway.


Read or execute the complete routing program

This Python reference was actually run during course preparation. It routes the six supplied metadata records, not real children or copyrighted works. Trace it on paper or use a teacher-approved environment. The program has no network access, AI model, file upload or publication function.

def triage(record):
    fields = {"private_data", "school_ok", "rights_checked", "credit_complete"}
    if type(record) is not dict or set(record) != fields:
        raise ValueError("supply exactly the four review flags")
    if any(type(record[k]) is not bool for k in fields):
        raise ValueError("review flags must be actual booleans")
    if record["private_data"]:
        return "STOP_PRIVATE_DATA"
    if not record["school_ok"]:
        return "HOLD_SCHOOL_REVIEW"
    if not record["rights_checked"]:
        return "HOLD_RIGHTS_REVIEW"
    if not record["credit_complete"]:
        return "HOLD_CREDIT"
    return "READY_FOR_TEACHER_REVIEW"


flags = [
    (False, True, True, True),
    (False, True, True, True),
    (False, True, False, True),
    (True, True, True, True),
    (False, True, True, False),
    (False, False, True, True),
]
keys = ("private_data", "school_ok", "rights_checked", "credit_complete")
for number, values in enumerate(flags, start=1):
    print("A" + str(number), triage(dict(zip(keys, values))))

Expected outputs in order: A1 and A2 READY_FOR_TEACHER_REVIEW; A3 HOLD_RIGHTS_REVIEW; A4 STOP_PRIVATE_DATA; A5 HOLD_CREDIT; A6 HOLD_SCHOOL_REVIEW. There are two ready-for-review records, three holds and one stop. The reference rejects missing/extra fields and non-boolean flag values rather than treating the word false as a true value.

Before reading the worked checks, explain why adding a creator name alone does not clear A3. Partial step: the rights flag remains false even if the credit flag is true. Then test the boundary between a false private-data flag and a true one with the other flags unchanged. The human reviewer needs the actual source and planned use, not only a green status.



Improve collaboration without hiding limits


Compare a convenient shortcut with the full routing rule

A simpler baseline says: if credit_complete is true, send the record to review; otherwise hold it. On A1 to A6 it sends A1, A2, A3, A4 and A6 onward and holds A5. Under our explicitly declared classroom routing policy, three of those onward routes miss a required stop or hold. Counting a name or credit field is therefore a poor substitute for examining privacy, school approval and rights evidence.

This is a six-scenario policy comparison, not measured copyright-detection accuracy. The target decisions are supplied by our own fictional policy, and the records do not prove the flags are true in the real world. An alternative triage system could have different categories or require more checks; explain its policy and evidence rather than pretending a toy rule is a universal legal standard.

The full program sends only A1 and A2 onward for human review. Retain A3 to A6 and the reason for each hold/stop in a safe project log; do not copy unnecessary personal content into that log. A corrected metadata record should preserve its version history. Changing a flag without doing the underlying review is not a correction.


Give people meaningful control and an honest account

For our poster, the editor should inspect the message, actual visual, source record, readability and contribution trail. The teacher can reject an output, request a different source, remove personal details or choose the human-only version. Human review needs enough information and authority to change the result; a compulsory final click is not enough.

A concise final note might say that the message was drafted and revised in the classroom, with a teacher-prepared assistance example discussed but no live pupil AI used. If a different real project uses an approved tool, record the actual tool, purpose and human changes under the school's policy. Do not copy our no-live-AI note into a project where that would be untrue.

Do not present a generated-looking image as documentary evidence of a real class event, or imply a real artist or organisation participated without evidence. Keep fictional and actual material distinguishable. A request to make a convincing poster does not justify false quotations, fabricated awards or a misleading claim that the library is officially endorsed by someone famous.

CodeAI: AI: Training Data & Bias. Discuss whose experiences or visual preferences might be missing from suggested content. The recording does not certify an output as fair, establish permission for its training material or approve a tool for a school. Evaluate the actual draft and its context, including stereotypes and accessibility.


State two limitations and improve the workflow

First, metadata can be incomplete or false. Mitigate this by having responsible staff inspect actual source pages, terms and intended use, retain the evidence and challenge inconsistent entries. A program cannot turn an unverified claim into a genuine licence. Even complete credits do not establish all rights or age suitability.

Second, a routing rule does not judge communication quality or identify every harmful or misleading element. Use a separate human audience/accessibility review, including a readable text equivalent and a way to request changes. These four teaching flags do not replace school safeguarding or data-protection processes or form a comprehensive institutional checklist.

One proposed future test would give learners a draft with ambiguous provenance and compare how reviewers request missing evidence. It has not been carried out here. Preserve that boundary in the evidence log. For a new audio or music project, redefine rights, performer contributions, attribution and accessibility needs; permission for one photograph is not permission to sample a song.

Before any separate live pupil AI use, staff must assess the specific tool, current age conditions, supervision, filtering, monitoring and school approval. Do not upload classmates' work, names, faces or private details merely because a tool is convenient. The present activity can be completed using the original text, six fictional records and credited media, without sending pupil work anywhere.

Keep a creative brief, earlier drafts, final choice and reasons, contribution/source notes, six-case routing log, baseline comparison and a teacher-review decision. These document learning and responsible collaboration; they do not establish independent legal clearance, a real AI product's quality or a separate statutory annual AI requirement.



Interactive Tasks


Quiz: Test Your Knowledge

What is the purpose of the fictional reading-corner poster? (Help learners choose a book and start a conversation) (!Collect home addresses) (!Prove that every AI suggestion is original) (!Give automatic legal clearance)



What is the assistance card in our version history? (An original teacher-prepared simulation for this lesson) (!A verified transcript from a named live chatbot) (!A real pupil private message) (!An automatically licensed quotation from a famous writer)



What does READY_FOR_TEACHER_REVIEW mean? (The record reaches a person who still checks the actual evidence) (!The program automatically publishes it) (!All possible legal rights have been proved) (!The draft must be accepted unchanged)



Why does A3 wait even though its credit flag is true? (The actual rights assessment remains unverified) (!A creator name always removes every restriction) (!The program has secretly published it) (!All web images are forbidden in every case)



Which priority applies to A4? (Stop unnecessary private information first) (!Ignore privacy because the other flags are true) (!Publish it with a source link) (!Convert the address into a public quiz answer)



Which record waits for a complete credit note? (A5) (!A1) (!A2) (!A4)



How many records reach teacher review under the full rule? (Two) (!Five) (!Six) (!Zero)



Which three onward baseline routes miss a required hold or stop? (A3 A4 and A6) (!A1 A2 and A5) (!A1 A3 and A5) (!A2 A4 and A5)



What is a useful reason for the final message revision? (It gives the stated audience two concrete actions) (!It proves a universal best wording) (!It hides the earlier contribution history) (!It guarantees a famous organisation endorsement)



What should happen before a different live pupil AI activity? (Staff assess the tool and school safeguards for that specific use) (!Pupils bypass age restrictions) (!All class work is uploaded by default) (!A green metadata flag replaces supervision)





Memory Game

Creative brief A purpose audience format and constraints
Contribution trail A record of who did what and what changed
Attribution An appropriate acknowledgement of a source or contribution
Permission A basis for the particular intended use
Triage Routing a record to an appropriate next check
Human review A person examines evidence and can reject the proposal





Drag and Drop

A creator name without verified rights Seek the actual source and responsible assessment
Unnecessary identifiable content Stop using it in this task and involve responsible staff
A draft suggestion changed by an editor Record both assistance and the actual human revision
An unapproved pupil service proposal Wait for school assessment instead of bypassing restrictions
A ready-for-review status Present the actual evidence to a person who can refuse





Crossword Puzzle

DRAFT What is an earlier version kept for revision?
SOURCE Where should an image record lead a reviewer?
CREDIT What acknowledges a creator or contribution?
REVIEW What should a person do before approving a final use?
PRIVACY What is at risk when unnecessary personal details are shared?
BRIEF What defines the creative purpose and audience?





LearningApps

Optional extra practice. Choose an appropriate activity with your teacher. If the external collection is unavailable, use the native activities and independent tasks in this course.


Cloze Text

Complete the explanation.
A creative

defines the audience and purpose. Keep earlier

and a contribution trail. Naming a creator gives

, not automatic permission. Inspect the original

. The program organises a

queue. A4 triggers a privacy

. Only

records reach teacher review. The shortcut checks credit

. A person must inspect the actual

. Our assistance card is a prepared

, not a live chatbot transcript.



Open Tasks


Easy

  1. Explain the audience: State what the poster should help a reader do.
  2. Read the version history: Separate learner-role drafting, simulated assistance and human revision.
  3. Trace a private-data stop: Explain A4 without copying any real private information.
  4. Read a source note: Identify creator, source, licence and change information in a credited image record.


Standard

  1. Trace all six records: Use the four flags in priority order and retain each reason.
  2. Compare the shortcut: Identify which onward baseline routes miss required checks under this fictional policy.
  3. Write a contribution note: Describe actual work and assistance honestly, preserving earlier versions.
  4. Improve the final poster: Propose a defensible revision for the stated audience with readable design and text alternative.


Advanced

  1. Challenge a green flag: Explain how false metadata can defeat a routing program and what evidence staff must inspect.
  2. Transfer to an audio project: Redefine rights, contribution, credit and accessibility needs without assuming photograph permission covers music.
  3. Design a new review test: Keep an ambiguous-provenance trial explicitly proposed until it is actually performed.
  4. Evaluate the technology choice: Compare a human-only workflow with approved assistance using the same audience criteria and honest evidence limits.



Learning Check

  1. Contribution and revision: Explain the three prepared text stages.
  2. Six routes: Apply the declared flag priorities.
  3. A shortcut comparison: Evaluate the credit-only baseline.
  4. Attribution and permission: Explain why a source name is insufficient.
  5. Limits and human authority: Give two limitations and mitigations.
  6. Transfer and an honest final note: Apply the process to a new creative project.


Worked checks and assessment guidance

  1. The learner-role line about journeys is V1. The tomorrow line is a teacher-prepared assistance simulation, not a real chatbot transcript. The editor-role final line gives two actions. Keep the drafts and explain the audience-based choice; do not invent a universal quality ranking or ownership finding.
  2. A1 and A2 reach READY_FOR_TEACHER_REVIEW. A3 waits for rights review, A4 stops private information, A5 waits for complete credit and A6 waits for school review. Totals are two onward records, three holds and one stop. None is automatically published or legally certified.
  3. The shortcut sends A1, A2, A3, A4 and A6 onward and holds A5. Against the explicitly stated classroom policy, A3 lacks rights review, A4 requires a privacy stop and A6 lacks school approval. This is policy routing on invented flags, not measured real copyright-detection accuracy.
  4. Record the actual source, creator, licence/version, intended use and changes as relevant. Naming a creator is not itself permission. Responsible staff check the actual material and rights; Creative Commons conditions and other rights may still matter. A copied licence badge or green flag cannot establish every permission.
  5. Metadata may be false or incomplete: inspect actual sources and retain evidence. Routing does not judge suitability, misleading claims or readability: use a separate audience/accessibility review. Give the teacher authority to reject or request changes. The proposed ambiguous-provenance reviewer trial remains unperformed.
  6. Redefine the audience, material, contributors and permissions for the new medium, including text alternatives or captions. Record actual assistance rather than copying an inaccurate no-AI note. Check school approval before live use and avoid unnecessary personal data. Use a human-only workflow when it meets the goal with fewer unresolved risks.


Learning evidence

Keep the creative brief, V1 and final choice with reasons, actual contribution/source notes, six-case triage log and credit-only baseline comparison. Distinguish two ready-for-review records, three holds and one stop. Explain two limitations and mitigations, one unperformed future test and a meaningful teacher-review decision. No legal clearance, real pupil AI use or automatic publication is established.


OER and Wikipedia

DfE Computing anchors the programming context; this AI Year 5 studio is editorial enrichment. DfE: generative AI in education was consulted for school responsibility and appropriate use. Creative Commons BY-SA 4.0 deed and CC considerations for licensors and licensees inform the distinction between a credit record and the actual scope of permission. These references are not a legal approval of the fictional poster. Reviewed 15 November 2026.

International books at Kent Library.jpg, Roc0ast3r, CC0. Unchanged original, used as library context; not a photograph of our fictional class.

https://www.youtube.com/watch?v=X-AWdfSFCHQ — CodeAI: How Chatbots and Large Language Models Work. Original recording embedded unchanged.

https://www.youtube.com/watch?v=x2mRoFNm22g — CodeAI: AI: Training Data & Bias. Original recording embedded unchanged. Teacher-supported context; no live pupil AI activity required.



Linked learning areas


aiMOOC Projects



Schulfach+




aiMOOCs



aiMOOC Projekte









MOOCwiki · Deutsch

Nach dem Lernen ist vor dem Lernen

Entdecke direkt den nächsten Lernkurs. Weitere Inhalte erscheinen, wenn Du weiter nach unten scrollst.

Zur MOOCwiki-Hauptseite

Mediathek

Suche verfeinern

Dein Lernplan bestimmt, was oben steht. Alle anderen Inhalte bleiben erreichbar. Seitenaufrufe beeinflussen die Auswahl nicht.

Inhalte werden geladen ...

Mediathek wird aus dem Wiki geladen ...