Die Druckversion wird nicht mehr unterstützt und kann Darstellungsfehler aufweisen. Bitte aktualisiere deine Browser-Lesezeichen und verwende stattdessen die Standard-Druckfunktion des Browsers.

aiMOOC

ABI - AI and education

QR-Code

ABI - AI and education

Mini-description: Compare documented opportunities, risks and educational approaches to generative AI. Reflect on academic integrity, independent thinking and the difference between producing a better answer and actually learning. This aiMOOC is designed for English in the Kursstufe in Baden-Württemberg, for a shared Basisfach / Leistungsfach course.

English first: Work from context, word families, visuals and English explanations. Use a monolingual dictionary before German support. German help is a safety net, not a full translation.

Level: mainly B2, with selected C1-oriented language and analysis. No artificial complexity.

Important Abi note, checked 03 October 2026: “AI and education” is not the official English Schwerpunkt for the Baden-Württemberg Abitur 2027 or 2028. The topic is used here as a current non-literary practice field for listening/viewing, reading, analysis, argumentation, speaking, writing and mediation. Always check the official documents for your own examination year.

Media entry: Look at the workshop photo for 30 seconds. What can you actually observe? What do you only infer? Then finish the sentence: “AI supports learning when …, but it may weaken learning when …”

Einleitung

Generative AI can produce text, images, audio and code from prompts. In education, this creates genuine opportunities: adaptive feedback, dialogue, language support, brainstorming and assistance with routine work. It also creates risks: unreliable output, bias, privacy problems, unequal access, unclear authorship and cognitive offloading — shifting mental work to a tool instead of practising it yourself.

The central question of this course is therefore not simply “Is AI good or bad for education?” A stronger question is: Under which conditions does AI support learning without replacing the thinking that learning requires?

The OECD’s 2026 Digital Education Outlook stresses an important distinction: general-purpose GenAI can improve task performance without necessarily producing learning gains. Pedagogically designed or carefully guided use can be more beneficial. UNESCO recommends a human-centred, age-appropriate and privacy-conscious approach. Baden-Württemberg’s ZSL likewise stresses teacher responsibility, data protection and transparent use.

OECD Digital Education Outlook 2026

UNESCO Guidance for Generative AI in Education and Research

ZSL Baden-Württemberg: FAQ zu KI

I-can goals

By the end of the course, you can:

Generative AI: explain in simple English what a generative model does and why fluent output is not automatically true.
Source criticism: distinguish observation, source-based claim, inference and value judgement.
Argumentation: compare documented opportunities, risks and educational approaches with evidence.
Academic integrity: explain authorship, attribution, verification and transparent AI use.
Independent thinking: protect your own reasoning through a think-first and verify-after process.
Text analysis: analyze structure, language, communicative strategies and visual elements in non-literary material.
Speaking: discuss competing educational choices and respond to counterarguments.
Writing: produce an independent, evidence-based text at B2 level, with selected C1-oriented precision.
Mediation: communicate relevant German information appropriately in English.
Abitur: use current operators and recognize which parts of this course train Abi-relevant skills.

BF and LF: one course, different depth

Area	Basisfach	Leistungsfach
Material	One core video plus one or two source cards	Several sources and independent material linking
Analysis	Selected aspects with clear structure aids	Broader and deeper analysis of purpose, evidence, language, structure and visuals
Argument	One clear claim-evidence-explanation chain	Synthesis of competing claims, limits and implications
Language	Secure B2 expression and useful sentence frames	B2 with selective C1-oriented precision and flexible register
AFB	Strong support for AFB I and II, selected AFB III	More independent AFB II and III work
Output in this course	Shorter, focused practice products	Longer, more autonomous and interconnected products

These practice lengths and pathways are didactic choices for this MOOC, not official examination word limits.

Help ladder

Before asking for a German translation, move through this order:

Context: What meaning fits the sentence?
Word family: Does a prefix, suffix or related word help?
Visual clue: Does a diagram, image or example clarify the idea?
English explanation: Paraphrase the word in simple English.
Monolingual dictionary: Check meaning, collocation and example.
German support: Use a short German explanation only if the meaning is still unclear.

Vorwissen aktivieren

Without using any AI tool, make a two-column note:

“AI can help learning when …”	“AI can harm learning when …”
Write three hypotheses.	Write three hypotheses.

Then mark each idea:

K = I know this from a source. H = It is my hypothesis. V = It is a value judgement.

This distinction will matter throughout the course.

How generative AI works: just enough for analysis

A language model learns statistical patterns from large amounts of data and generates output step by step. It can produce highly plausible language without independently guaranteeing that a statement, quotation or source is correct. That is why verification remains a human task.

Visual task: Follow the arrows. Explain in 60 seconds why “trained on data” does not mean “stores a perfect database of verified facts”.

A model diagram simplifies reality. Use it to build a concept, not to claim that every generative system is trained in exactly the same way.

Listening and Viewing

OECD, 2026: “How is Generative AI impacting education?”

First viewing: gist

Watch once without stopping.

Gist: What is the central distinction made in the video?
Key message: Complete: “Better performance does not automatically mean …”
Purpose: Is the video mainly advertising a product, explaining research findings, entertaining viewers or giving examination advice? Support your answer.

Second viewing: evidence grid

Claim in the video	Evidence or explanation	Educational implication
General-purpose AI may improve performance.	What reason is given?	What should a learner or teacher do differently?
Pedagogical intent matters.	What kind of use is contrasted with simple outsourcing?	What would this look like in an English lesson?

BF: Complete one row carefully and present it orally.

LF: Complete both rows, then compare the video’s message with the OECD report page. Note one point that is supported, one limitation and one question that remains open.

A positioned perspective

Sal Khan’s TED talk presents a strongly opportunity-oriented argument about AI tutoring and teaching assistance. Treat it as a positioned source, not as neutral proof.

Claim: Identify one major promise.
Evidence: What kind of evidence or demonstration is used?
Perspective: Why does the speaker’s role matter when evaluating the argument?
Comparison: Compare the talk with the OECD distinction between task performance and learning.

Optional viewing: You may use teacher-provided excerpts or the public video. No film upload, screen recording or external AI tool is required.

Reading: source cards, not text walls

The following cards are concise paraphrases of the linked sources. Read the original source selectively when you need evidence. Do not copy full protected texts into your work.

Source A: OECD 2026

Documented message: GenAI can support learning when its use follows clear teaching principles. Simply outsourcing cognitive work can improve an immediate product without creating equivalent learning gains. The OECD therefore emphasizes independent thinking, foundational skills and selective, purposeful use.

Analysis prompt: Which educational problem is the OECD trying to prevent?

Open the OECD source

Source B: UNESCO

Documented message: UNESCO recommends a human-centred approach to GenAI in education. Its guidance emphasizes privacy, age-appropriateness, ethical validation, inclusion and pedagogical design.

Analysis prompt: Which of these principles protects learning, and which mainly protects rights or safety? Where do they overlap?

Open the UNESCO source

Source C: ZSL Baden-Württemberg

Documented message: AI systems may support tutoring, feedback and differentiated materials, but teachers remain responsible for assessment. Personal data should only be entered into systems approved for that purpose, and AI-supported feedback must meet transparency and data-protection requirements.

Analysis prompt: Why is “technically possible” not the same as “educationally or legally appropriate”?

Open the ZSL source

Context vocabulary

Word	English first	Word family / clue	German fallback
scaffold	temporary support that helps a learner perform a task	scaffold, scaffolding	Lernstütze
cognitive offloading	shifting mental work from yourself to a tool	cognition, cognitive	Auslagern geistiger Arbeit
hallucination	plausible-looking model output that is false or unsupported	hallucinate, hallucinatory	erfundene oder falsche KI-Ausgabe
verification	checking whether a claim is supported by reliable evidence	verify, verifiable	Überprüfung
bias	a systematic tendency that can distort representation or decisions	biased, unbiased	Verzerrung
attribution	stating where an idea, quotation or contribution came from	attribute, attributable	Quellen- oder Urheberangabe
authorship	responsibility for creating and owning a piece of work	author	Urheberschaft
integrity	acting according to agreed ethical and academic rules	integrity, integrated	Integrität
equity	fairness that takes different needs and barriers into account	equitable	Chancengerechtigkeit
overreliance	depending on a tool more than is useful or safe	rely, reliable, reliance	Überabhängigkeit
transparency	making processes, sources or tool use visible and understandable	transparent	Transparenz
independent thinking	forming and checking your own reasoning instead of delegating it	independence, independently	eigenständiges Denken

Opportunities, risks and educational approaches

Opportunity	Risk or limitation	Educational approach
Fast feedback	Feedback may be wrong, generic or overtrusted.	Compare feedback with criteria, evidence and your own revision decision.
Dialogue and tutoring	The learner may become passive or accept suggestions too quickly.	Use questions, hints and scaffolds rather than answer replacement.
Language support	A polished output may hide gaps in the learner’s language competence.	Draft independently first; use support for noticing and revision.
Brainstorming	Generated ideas can narrow originality or introduce unsupported claims.	Generate your own ideas first; compare, reject, combine and verify.
Differentiation	Access and quality may be unequal.	Provide no-AI alternatives and teacher-designed support.
Teacher assistance	Personal data, copyright and automated judgement create risks.	Use approved systems, minimize data and keep human responsibility.

Core distinction: An educationally useful tool should not merely make the final product look better. It should help the learner practise, understand, reflect and become more independent.

Media literacy: what does an image prove?

Datei:Participants at the AI panel of the EduWiki Hub 05.jpg

This photograph documents that people discussed AI and education at an EduWiki event. It does not prove that any particular claim about AI is true.

Use the four-step visual routine:

Describe only what is visible.
Contextualize source, date and setting.
Interpret possible meanings.
Evaluate what the image can and cannot support as evidence.

Compare the two photographs: How do composition, setting and captions influence your impression of expertise, participation or authority? Do not infer more than the images can support.

Academic integrity and independent thinking

Academic integrity is not only about avoiding cheating. It also means that your submitted work represents your own assessable thinking, that sources are traceable and that permitted assistance is disclosed according to school rules.

Use the Think → Assist → Verify → Rewrite → Disclose protocol:

Think: Produce an initial idea, outline or answer yourself.
Assist: If AI use is permitted, use it for a limited function such as questions, feedback or alternative wording.
Verify: Check factual claims, quotations and sources against reliable material.
Rewrite: Decide independently what to keep, reject or change.
Disclose: State permitted AI assistance in the form required by your teacher or institution.

Never submit an AI-generated replacement essay as your own work. Never invent sources. Never treat a generated citation as real until you have opened and checked it.

Privacy rule for this MOOC: Use fictional data. Recordings are voluntary and may remain private. No task requires you to upload personal work, personal data or recordings. No task requires access to an external AI system.

A no-AI integrity exercise

Evaluate these fictional statements:

Fictional statement	Diagnose the problem	Improve it
“AI gives students facts instantly, therefore students learn more.”	Does the conclusion follow from the premise?	Add a condition, limitation or source.
“A chatbot cited a study, so the study exists.”	What needs verification?	State a reliable checking procedure.
“My essay is mine because I typed the prompt.”	Which aspect of authorship is missing?	Explain what independent contribution should be visible.

Mark your revised sentences with:

S = source-supported claim I = inference J = judgement

Fictional data for visualisation

The following dataset is invented for practice. It is not research evidence.

A fictional class of 40 learners is asked what they would do before using optional AI support for a writing task.

Response	Fictional number of learners
Write an independent first draft	18
Make an outline first	10
Ask AI before thinking independently	7
Use no AI	5

Task: Turn the data into a bar chart on paper or with a locally approved tool. Then write three sentences: one accurate description, one cautious interpretation and one statement that the data cannot justify.

Analyse

For non-literary texts and media, use the chain:

purpose → structure → evidence → language → communicative strategies → visual design → effect → evaluation

Useful questions:

Purpose: What does the author or institution want the audience to understand or do?
Structure: How is the argument sequenced?
Evidence: Are claims supported by research, examples, statistics or authority?
Language: Which lexical fields, register, tone or syntactic patterns matter?
Communicative strategies: How do direct address, examples, contrasts or qualifications shape the message?
Visuals: How do composition, scale, captions, diagrams or statistics affect meaning?
Evaluation: How strong is the claim given the evidence and its limitations?

Operator focus from examination year 2027: In the IQB operator stock, analyze / examine requires detailed description and explanation supported by textual evidence and related to relevant aspects of language and form. Assess / evaluate requires a well-founded judgement supported by evidence and examples.

IQB operator stock for English, valid from examination year 2027

Speaking

Motion for discussion: “Schools should allow generative AI whenever the use is transparent.”

Do not vote immediately. First build both sides.

Support the motion	Challenge the motion
Name one documented opportunity.	Name one documented risk.
Give one condition for responsible use.	Give one situation where AI should not replace independent work.
Use one source.	Use one source.

BF support: Use the frame: “One benefit is … . However, this depends on … . The OECD/UNESCO/ZSL source suggests … . Therefore … .”

LF extension: Synthesize two sources, identify a tension between them and respond to a counterargument before giving your judgement.

Recording: Optional only. You may practise live with a partner or record privately. No upload is required.

Writing

BF pathway

Write a focused article for a fictional school website:

Task: “Explain one opportunity and one risk of generative AI in learning and propose two rules that protect independent thinking.”

Recommended practice length: about 250–300 words.

Structure support:

Opening: define the issue and your focus.
Opportunity: claim, evidence, explanation.
Risk: claim, evidence, explanation.
Approach: two realistic rules.
Closing: answer the central question.

LF pathway

Write an evidence-based response for a fictional education magazine:

Task: “Assess to what extent generative AI can support learning without weakening independent thinking.”

Recommended practice length: about 400–500 words.

Use at least two of the course sources. Distinguish documented findings from your inference. Include one meaningful counterargument and one limitation of the available evidence.

No replacement essays: The purpose is your analysis and writing. If AI use is permitted by your teacher, limit it to a declared support function and keep your own planning, source check and final decisions visible.

Mediation

German source card for mediation — written for this MOOC:

„KI kann im Unterricht beim Üben, bei Rückmeldungen und bei der Differenzierung unterstützen. Dabei bleiben Lehrkräfte für die Bewertung verantwortlich. Personenbezogene Daten dürfen nicht unkontrolliert in KI-Systeme eingegeben werden. Lernende sollen erkennen können, wann KI eingesetzt wurde. Unterricht muss außerdem sicherstellen, dass eigenständiges Denken und fachliche Kompetenzen nicht durch bloßes Auslagern von Aufgaben ersetzt werden.“

Scenario: An English-speaking exchange student joins your school’s student council. Mediate the relevant information in an email explaining how responsible AI use could be organized.

BF: Select the three most relevant points and use clear B2 language.

LF: Prioritize information for the situation, make links between privacy, transparency and learning, and adapt register independently.

Do not translate sentence by sentence.

Feedback and error analysis

Use this short feedback code:

Code	Check
C	Is the claim clear?
E	Is there relevant evidence?
L	Is the logical link explained?
S	Is the source identifiable and checked?
I	Is inference clearly distinguished from fact?
V	Is vocabulary precise enough?
A	Is authorship and permitted assistance transparent?

Error-analysis routine: Choose one weak paragraph. Underline the claim, box the evidence, circle the explanation and mark unsupported leaps with “?”. Revise only after you can name the problem.

Transfer

Design a one-page AI use card for a fictional English course. It must answer:

Learning goal: What skill should remain human and assessable?
Permitted support: What may AI help with?
Boundary: What would replace the target competence?
Verification: What must be checked?
Disclosure: How is assistance documented?
Privacy: What data must not be uploaded?

Then test your card against two scenarios: vocabulary practice and an assessed analytical essay. A strong policy may treat the two situations differently.

Abi-Check Baden-Württemberg

Status checked: 03 October 2026. Use the official documents for your own examination year as the final authority.

Check	What matters
Topic status	“AI and education” is not the official 2027/2028 English Schwerpunktthema. For both years, the listed Schwerpunkt is “On the Move: Migration and Cross-Cultural Encounters”.
Skills link	This MOOC nevertheless practises Abi-relevant work with non-literary texts, listening, analysis, visuals, argumentation, speaking and mediation.
LF written examination 2028	The official framework states about 255 minutes including selection time: listening about 30 minutes, then a 15-minute break, then 225 minutes for writing.
LF writing	The official structure includes an aspect-focused summary, analysis, and a choice involving personal response or creative writing. Non-literary analysis may include structure, language, communicative strategies and discontinuous visual material.
LF speaking	The examination framework includes a communication examination.
BF	BF and LF share the Schwerpunkt in an appropriately reduced scope. Across the course, BF assessment covers listening, reading, analysis, mediation and different writing types; the Abitur examination in the Basisfach is oral.
Level	The Baden-Württemberg curriculum states B2 at the end of the Kursstufe, partly C1.
Operators	The revised IQB operator stock applies from examination year 2027.

Official sources:

Facherlass Abitur 2027

Facherlass Abitur 2028

Bildungsplan: Basisfach und Leistungsfach

IQB: Begleitende Dokumente

Folgekurs

Continue with AI, media literacy and society. Transfer the same source-check routine to synthetic media, online information and public communication: observe, contextualize, verify, analyze and only then evaluate.

Interaktive Aufgaben

Quiz: Teste Dein Wissen

What distinction is central to the OECD Digital Education Outlook 2026? (Better task performance does not automatically mean better learning) (!Any use of AI automatically improves long-term learning) (!AI can only be useful outside education) (!Learning and task performance always mean the same thing)

What does cognitive offloading mean in this course? (Shifting mental work from yourself to a tool) (!Learning a new vocabulary list) (!Checking two independent sources) (!Explaining an argument to a partner)

Which use is an educational opportunity of generative AI? (Providing guided feedback and tutoring support) (!Guaranteeing that every generated fact is true) (!Replacing all independent practice) (!Removing the need for teachers)

Which action best supports academic integrity? (Disclosing permitted assistance and verifying sources) (!Submitting generated work as your own) (!Inventing a source when none is available) (!Assuming fluent language proves factual accuracy)

What is a hallucination in the context of generative AI? (A plausible but false or unsupported output) (!A verified quotation from a source) (!A teacher-approved assessment criterion) (!A monolingual dictionary definition)

Which privacy principle fits this course? (Do not enter personal data into unapproved AI systems) (!Upload classmates' work for comparison) (!Share private recordings by default) (!Use real personal data whenever an example needs detail)

What does the 2028 Baden-Württemberg LF written English examination include? (Listening and writing) (!Only a grammar test) (!Only mediation) (!Only a prepared speech)

How do Basisfach and Leistungsfach mainly differ according to the curriculum? (In complexity breadth depth and differentiation) (!In using completely unrelated school subjects) (!In having different target languages) (!In avoiding non-literary texts in the Basisfach)

Which statement about AI and education is correct for the 2027 and 2028 English Abitur in Baden-Württemberg? (It is a skills practice topic here but not the official Schwerpunkt) (!It replaces all official Pflichtwerke) (!It is the only official examination theme) (!It removes the need to check the annual Facherlass)

What is required when you analyze or examine a text under the revised IQB operator guidance? (Explain relevant aspects in detail and support them with evidence) (!Give only your personal opinion) (!Retell the text without analysis) (!List vocabulary without linking it to meaning)

Memory

scaffold	temporary learning support
verification	checking a claim against reliable evidence
attribution	stating where an idea or contribution came from
cognitive offloading	shifting mental work to a tool
equity	fairness that considers different needs and barriers
authorship	responsibility for creating a piece of work

Drag and Drop

Ordne die richtigen Begriffe zu.	Thema
Think first	independent baseline
Source check	factual verification
Process log	transparent authorship
Private practice	voluntary speaking rehearsal
Teacher scaffold	structured learning support

Kreuzworträtsel

Scaffold	What is temporary support that helps a learner complete a task?
Bias	What word describes a systematic tendency that can distort representation?
Verify	Which verb means to check whether a claim is supported?
Authorship	What term describes responsibility for creating a piece of work?
Equity	What word describes fairness that considers different needs and barriers?
Feedback	What term describes information used to improve a performance or product?

LearningApps

Lückentext

Generative AI can improve task

without automatically improving learning. Shifting too much mental work to a tool is called cognitive

. A learner should

factual claims before using them. Transparent acknowledgement of outside help supports academic

. UNESCO recommends a human-centred and privacy-aware educational

. The Baden-Württemberg curriculum expects B2 and in parts

by the end of the Kursstufe. The revised IQB operator stock applies from examination year

. In this MOOC, AI and education is a practice topic rather than the official Abitur

.

Offene Aufgaben

Leicht

Vocabulary map: Create a one-page visual map linking six course terms with arrows and short English definitions.
Source check: Choose one factual claim from a course source and document how you verified it using the original page.
Media caption: Write an accurate English caption for one Commons image and add one sentence explaining what the image cannot prove.
Speaking rehearsal: Give a 90-second explanation of the difference between performance and learning; recording is optional and may remain private.

Standard

Evidence grid: Compare the OECD video with one written source using claim, evidence, limitation and implication.
Integrity guide: Design a five-rule student guide for transparent and independent AI-supported learning.
Mediation task: Mediate the German source card into an English email for an exchange student without translating sentence by sentence.
Data visualisation: Turn the fictional class data into a chart and write a cautious interpretation plus one limitation.

Schwer

Comparative analysis: Compare the OECD and TED videos in terms of purpose, evidence, tone and communicative strategy.
Policy design: Draft a one-page AI-use policy for a fictional English course and justify different rules for practice and assessment.
Synthesis writing: Assess to what extent GenAI can support learning without weakening independent thinking, using at least two documented sources.
Research audit: Select three claims about AI and education from public sources, classify their evidence, identify uncertainty and produce a verified mini-dossier.

Lernkontrolle

Case 1: A learner produces an excellent AI-assisted essay but cannot explain its argument orally. Analyze what this tells you about performance, learning and assessable authorship.
Case 2: Redesign an AI activity that currently gives students complete answers so that it instead scaffolds independent thinking.
Case 3: Compare a claim from an advocacy talk with a claim from an institutional research source. Explain how source role, evidence and purpose affect your judgement.
Case 4: A teacher wants to upload student essays to an unapproved chatbot for feedback. Explain the educational and privacy issues and propose a safer alternative.
Case 5: Create two versions of the same analysis task, one appropriate for BF and one for LF, without making the BF task trivial or the LF task artificially difficult.
Case 6: Take an unfamiliar chart or short non-literary text about educational technology and plan how you would summarize, analyze and evaluate it using the current operator logic.

Lernnachweis

For a successful learning record, show that you can:

explain the difference between task performance and learning;
compare at least one documented opportunity, risk and educational approach;
distinguish fact, inference and judgement;
analyze purpose, structure, language, evidence and visual elements;
use sources transparently and verify factual claims;
explain academic integrity, authorship and independent thinking;
produce an independent spoken or written response appropriate to BF or LF;
reflect on one error or weak reasoning step and improve it;
apply the Think → Assist → Verify → Rewrite → Disclose protocol;
connect the course skills to the current Baden-Württemberg Abi framework without claiming that AI and education is an official 2027/2028 Schwerpunkt.

OERs zum Thema

Further open or public reference material:

Wikimedia Commons: Artificial intelligence

Wikimedia Commons: AI and Wikipedia in Education

Wikipedia: Künstliche Intelligenz in der Hochschullehre

Quellen und Mediennachweise

The course uses source summaries rather than reproducing complete protected texts. Embedded videos remain on their official platforms. Wikimedia Commons files are linked through their Commons filenames and retain their licence information on the file pages.

Verknüpfte Lernbereiche

ABI - AI and education

aiMOOC-Projekte

Frag mich zum aktuellen Lerninhalt oder wähle eine kurze Hilfe.

Textmodus bereit

Sprache