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 LF - AI debate challenge

QR-Code

ABI LF - AI debate challenge

Mini-description: Lead a structured, evidence-based debate on one clearly defined AI question. Represent competing positions fairly, test claims against evidence, and communicate at Leistungsfach level.

Target group: Englisch als fortgeführte Fremdsprache, Kursstufe Baden-Württemberg, Leistungsfach. English first; German support comes only after context, word formation, visualisation, an English explanation and dictionary work.

Einleitung

Debate motion: This House believes that upper-secondary schools should allow students to use generative AI for homework and formative writing, provided every use is disclosed and independently verified.

The motion is deliberately narrow. It concerns homework and formative writing, not the Abitur examination and not hidden AI-written replacement essays. Disclosed means that AI assistance is named. Independently verified means that factual claims are checked against reliable sources rather than accepted from the same AI system.

Your challenge is not to “win at all costs”. Your task is to construct the strongest evidence-based case you can, represent the opposing case fairly, rebut precisely and state the limits of your evidence.

Media entry: Watch selected passages from the Oxford Union debate. Focus on structure, evidence, rebuttal and concessions, not on copying speakers' positions.

Safety, privacy and academic integrity

No external AI tool is required for this aiMOOC. Do not enter personal data into external systems. Any recording task is voluntary and may stay private on your own device. Do not upload recordings, images or written work unless your teacher explicitly asks you to use an approved school platform. If AI is permitted in a learning task, document its use and verify its claims. Your final analysis and writing must remain your own.

If a task uses a class poll or sample data, those data are marked as fictional and are for analysis practice only.

I-can goals

I can ...	Leistungsfach focus
narrow a broad AI topic into a debatable motion	precision and conceptual control
distinguish claim, evidence, warrant, counterclaim, rebuttal and limitation	AFB II/III reasoning
compare studies without overgeneralising from one subject, age group or setting	source criticism
analyse structure, tone, register, examples, statistics and communicative strategy	non-fiction analysis
build a five-minute evidence-based monologue and sustain a ten-minute dialogue	communication practice
write an independent, source-based argument without outsourcing the thinking	written exam transfer
mediate relevant information selectively and appropriately for an audience	optional mediation transfer

Vorwissen aktivieren

Answer in English before opening a dictionary.

Evidence: What makes information usable as evidence rather than merely an opinion?
Generative AI: Name one capability and one limitation.
Transfer: Why might evidence from mathematics not automatically apply to English writing?
Fairness: What would a fair summary of a position you disagree with sound like?

Diagnostic rule: If you cannot explain a term, first infer from context, then inspect its word family, then use an English learner's dictionary. Use a bilingual dictionary only afterwards.

Context vocabulary and word formation

Word	Word family	English explanation	Useful collocation
evidence	evident, evidential	information used to support or challenge a claim	strong / limited evidence
verify	verification, verifiable	check whether a claim is accurate	independently verify a claim
disclose	disclosure	make relevant assistance or information visible	disclose AI use
valid	validity, validate	well-founded for the purpose claimed	external validity
reliable	reliability, rely	consistent or dependable enough for a purpose	reliable source
bias	biased	a systematic tendency that can distort judgement or output	detect bias
scaffold	scaffolding	temporary support that helps a learner perform a difficult task	instructional scaffolding
overreliance	rely, reliance	depending on a tool so much that independent performance may weaken	avoid overreliance
agency	agent, agentic	a person's capacity to make informed choices and act independently	learner agency
equitable	equity, inequity	fair in access, opportunity and treatment	equitable access
rebut	rebuttal	answer a counterargument with reasons or evidence	evidence-based rebuttal
substantiate	substantiation, substantiated	support a claim with adequate evidence	substantiate an argument

Dictionary challenge: Add pronunciation, one example sentence and one extra collocation for six words.

Deutsch-Nothilfe erst zuletzt: Wenn Kontext, Wortfamilie, englische Erklärung und Wörterbuch nicht reichen, notiere eine knappe deutsche Arbeitsbedeutung. Übersetze nicht den gesamten Text.

Listening and Viewing

Viewpoint A: potential benefits

Sal Khan argues for educational uses of AI tutoring and teaching support. Treat the talk as an argumentative source, not as experimental proof.

Listening for claims: Note the central claim in one sentence.
Examples: Identify two examples used to make the claim plausible.
Source type: Explain what the talk can show and what it cannot prove.
Language analysis: Track one rhetorical strategy and its effect.

Viewpoint B: risks and governance

Gary Marcus argues that unreliable AI systems can create social risks and calls for stronger safeguards. Again, separate expert argument from empirical study.

Listening for counterclaims: State the main risk claim.
Evidence audit: Mark each support item as example, expert judgement, data or reference to research.
Rhetoric: Analyse tone, urgency and audience appeal.
Fair comparison: Write one sentence both speakers could plausibly agree with.

Reading: evidence pack

The following cards deliberately mix experimental evidence and policy guidance. Do not treat all source types as if they prove the same thing.

Evidence card A: high-school mathematics

A 2025 PNAS field experiment examined generative-AI tutoring in high-school mathematics. A standard chat-style condition improved performance while AI was available, but later unassisted performance was worse than for students who had not used that unrestricted condition. A more strongly scaffolded tutor was designed to give hints rather than simply provide answers.

Use carefully: This is strong causal evidence for its tested setting, but it is mathematics, not English writing. Do not generalise beyond the population, task and tool design without qualification.

PNAS: Generative AI without guardrails can harm learning

Evidence card B: university physics

A 2025 open-access randomized controlled trial with 194 students in a Harvard physics course compared a research-based AI tutor with in-class active learning for the same content. In that setting, students using the designed AI tutor learned more in less time and reported greater engagement and motivation.

Use carefully: The participants were university physics students and the tutor was purpose-built. This does not establish that unrestricted chatbots improve upper-secondary English writing.

Scientific Reports: AI tutoring outperforms in-class active learning

Evidence card C: human-centred guidance

UNESCO's guidance on generative AI in education stresses human agency, privacy, equity, age-appropriate use and pedagogical validation.

Use carefully: Guidance can frame ethical and policy criteria. It is not a randomized learning-effect study.

UNESCO: Guidance for generative AI in education and research

Evidence card D: digital literacy and disinformation

The European Commission's 2026-updated guidance for educators links generative AI with digital literacy, critical thinking and the challenge of misleading information.

Use carefully: Use this source for educational goals and risk awareness, not as proof that one classroom policy causes a particular learning outcome.

European Education Area: digital literacy and disinformation

Evidence quality: analyse before you argue

Question	What a strong answer does
What is the claim?	states it precisely and neutrally
What kind of evidence is offered?	distinguishes RCT, observational data, policy guidance, expert argument and anecdote
How direct is the evidence?	checks population, subject, task, tool and outcome
What is the warrant?	explains why the evidence supports the claim
What is the limitation?	names what cannot safely be inferred
What would change your mind?	identifies disconfirming evidence

High-level move: A study can be methodologically strong and still have limited transfer to your exact motion.

Visualisation: fictional class snapshot

Fictional data for analysis practice only — not evidence about real students.

Position before the debate	Fictional responses
allow AI with disclosure and verification	12
allow AI with few restrictions	3
do not allow AI for formative writing	9

Analyse the table in three stages: describe the pattern, interpret one possible reason, then limit the interpretation. Do not claim that this fictional class represents Baden-Württemberg or students in general.

Debate architecture

Move	Sentence frame
Claim	Our position is that ...
Evidence	This is supported by ...
Warrant	This matters because ...
Limitation	However, this evidence is limited to ...
Counterclaim	A strong objection is that ...
Rebuttal	That objection is important, but ...
Concession	We accept that ... while maintaining that ...
Synthesis	The evidence suggests a conditional position: ...

Fairness rule: Before rebutting, summarise the opposing argument in a form that the other team accepts as accurate.

Speaking challenge

Preparation

Narrow the motion: Keep the debate about homework and formative writing, not final examinations.
Build two cases: Prepare the strongest pro and con case before team roles are assigned.
Select evidence: Use at least one experimental study and one ethical or policy source.
Add limitations: Every major empirical claim must carry a transfer limitation.
Prepare questions: Write three questions that could genuinely test your own side.

Monologue practice

Deliver a structured five-minute position statement without reading a full script. Use notes containing keywords, evidence references and signposts only.

Support: Use three blocks: claim – evidence – limitation.

Core: Use two evidence cards, one concession and one rebuttal.

Stretch: Compare the evidential value of two source types and explain why they justify different levels of certainty.

Dialogue practice

Hold a ten-minute evidence-focused dialogue. Your goal is not rapid-fire contradiction. Ask for clarification, test relevance, challenge transfer, concede valid points and reformulate when new evidence changes your position.

Useful gambits: Could you clarify what outcome you mean? – How directly does that study apply here? – I accept that point, but the limitation is ... – What evidence would distinguish between these explanations?

Writing

Do not write a replacement essay with AI. Build the reasoning yourself.

Task: Write a 350–450 word evidence-based comment on the debate motion. Use at least two source cards. Your text must include one counterclaim, one rebuttal and one explicit limitation.

Planning frame before prose:

Element	Notes
precise thesis
strongest evidence for
strongest evidence against
transfer limitation
concession
final conditional judgement

LF stretch: Make the conclusion conditional rather than absolute by specifying the circumstances under which your judgement would change.

Optional mediation

Read the German-language version of the European Commission's digital-literacy guidance. Then brief an English-speaking exchange student in 120–150 words on the two points most relevant to an AI debate at school. Select; do not translate sentence by sentence.

Europäische Bildungsplattform – deutsche Fassung

Feedback and error analysis

Peer feedback

Criterion	Feedback prompt
precision	Which term or claim needs narrowing?
evidence	Which claim is best supported, and why?
warrant	Where is the reasoning link missing?
limitation	Where is transfer too broad?
interaction	Where did the speaker respond directly to the partner?
language	Which phrase improved register or cohesion?

Feedback must refer to observable performance. Suggest a better formulation rather than repeating an error.

Diagnose these faulty moves

Overgeneralisation: “The physics study proves AI improves learning in every subject.”
Source-type confusion: “A TED talk proves the causal effect.”
Unsupported leap: “AI can make practice faster, therefore schools must allow unrestricted use.”
Cherry-picking: Using only the study that supports your assigned side.
False balance: Treating a weak anecdote as equal to a relevant randomized study merely because it supports the other side.
Vague attribution: “Studies say ...” without naming what was studied.

Rewrite each move so that the claim becomes accurate, limited and source-aware.

Differentiation

Dimension	Support	Core	Stretch
language	sentence frames and vocabulary table	independent academic phrasing	nuanced qualification and register shifts
scope	two evidence cards	all four cards	one additional teacher-approved peer-reviewed or official source
pace	90-second micro-turns	full monologue and dialogue	rapid response plus synthesis
analysis	identify claim and evidence	evaluate relevance and limitations	compare methods and external validity
help	English definition and dictionary first	minimal prompts	no sentence frames during final round

Abi-Check Baden-Württemberg 2027

Status checked: 3 October 2026. For the 2027 Leistungsfach, the official Facherlass still refers to the Bildungsplan 2016 for the examined cohort. The mandatory focus topic is On the Move: Migration and Cross-Cultural Encounters, together with the prescribed works and the two thematic fields on politics, culture and society in the USA and the United Kingdom. AI is not the 2027 prescribed focus topic. This aiMOOC therefore trains transferable LF competences rather than replacing the official content.

Kultusministerium Baden-Württemberg: Facherlass Abitur 2027

Written examination: what this challenge trains

According to the 2027 Facherlass, the written LF examination is about 255 minutes including selection time: roughly 30 minutes listening, a 15-minute break, then 225 minutes for the writing part. In Part II, students choose between a literary and a non-literary text basis. The sequence includes an aspect-based summary, analysis, and then a choice between a personal statement and a creative writing task. Discontinuous materials such as images, diagrams or statistics may also be included.

This challenge therefore trains: listening for claims, aspect-based summary, analysis of line of argument and communicative strategies, visual-data analysis, and source-based personal comment.

Leistungsfach depth

The 2027 Facherlass states that LF course assessments should cover AFB I, II and III, with AFB II and III more strongly emphasised than AFB I. For this challenge, that means less retelling and more analysis, evaluation, source comparison, qualification and transfer.

Communication examination: keep it distinct

The 2027 Facherlass treats the communication examination separately and refers its detailed design and assessment to the current Beurteilungs- und Korrekturrichtlinienerlass for modern foreign languages. The Landesbildungsserver's practice guidance describes 15 minutes of preparation, about five minutes of monologic speaking and about ten minutes of dialogic speaking. Use this aiMOOC's five-minute monologue and ten-minute dialogue as targeted practice, but check the current official assessment documents before a graded simulation.

Landesbildungsserver: Kommunikationsprüfung Bildungsplan Englisch Leistungsfach 11/12

Transfer

Use the same evidence routine with an official 2027 topic.

On the Move: Test a claim about migration using two independent evidence types.
USA: Analyse how a source frames freedom, equality or identity.
United Kingdom: Compare two perspectives on social or cultural change.
Arrival: Discuss only teacher-provided, legally accessible material; do not reproduce or upload film sequences.
Short stories: Work from licensed classroom copies; do not paste full required texts into the MOOC.

Media literacy checkpoint

Before trusting an AI-supported claim, ask: Who is responsible for it? What source can be opened? Is the evidence current? Is the quotation accurate? Does the source actually support the claim? What population and task were studied? Can an independent source confirm it?

Interaktive Aufgaben

Quiz: Teste Dein Wissen

Which statement best defines the motion in this aiMOOC? (Generative AI may be used for homework and formative writing if use is disclosed and independently verified) (!Generative AI should replace all English writing tasks) (!Generative AI is permitted in the Abitur examination) (!All AI systems should be banned from education)

What is a warrant in an argument? (The reasoning that explains why evidence supports a claim) (!A list of vocabulary items) (!A source title without analysis) (!A personal preference stated strongly)

Why must the PNAS mathematics study be transferred cautiously to English writing? (It tested a different subject, task and setting) (!It contains no learners) (!It is a fictional classroom survey) (!It is a dictionary entry)

What can a policy guidance document do best in this debate? (Provide ethical and policy criteria) (!Prove a causal learning effect by itself) (!Replace all empirical studies) (!Guarantee that every AI output is correct)

Which source type can most directly support a causal claim when well designed? (A randomized controlled trial) (!An unsupported social-media comment) (!A slogan) (!A personal anecdote)

What does independent verification require? (Checking an AI-supported claim against reliable external sources) (!Asking the same AI the same question twice) (!Using only the first search result) (!Trusting a fluent answer automatically)

What is a fair rebuttal? (A response that represents the opposing claim accurately before challenging it) (!A personal attack on the speaker) (!A louder repetition of the original claim) (!A refusal to acknowledge evidence)

What is the role of a limitation statement? (It marks the boundary of what the evidence can justify) (!It makes every argument invalid) (!It removes the need for evidence) (!It turns an opinion into a fact)

Which skill is especially important at Leistungsfach level? (Connecting and evaluating evidence from different materials) (!Copying one source without comment) (!Avoiding all counterarguments) (!Reducing every issue to one sentence)

Is AI the prescribed English LF focus topic for Baden-Württemberg Abitur 2027? (No, this challenge trains transferable competences) (!Yes, AI replaces all prescribed works) (!Yes, only AI is examined) (!No, because English has no prescribed content)

Memory

Claim	Debatable position
Evidence	Supporting information
Warrant	Reasoning link
Rebuttal	Response to a counterargument
Limitation	Boundary of inference
Guardrail	Designed restriction
Verification	Independent checking

Drag and Drop

Ordne die richtigen Begriffe zu.	Thema
Narrow the motion	Define exactly what is being judged
Present evidence	Support a claim with relevant information
State the warrant	Explain why the evidence matters
Add a limitation	Mark the boundary of the inference
Rebut fairly	Answer the strongest counterargument accurately

Kreuzworträtsel

Evidence	What supports or challenges a claim?
Warrant	What links evidence logically to a claim?
Rebuttal	What answers a counterargument?
Bias	What can systematically distort judgement?
Validity	What describes whether an inference is well founded for its purpose?
Disclosure	What makes relevant AI assistance visible?

LearningApps

Lückentext

Offene Aufgaben

Leicht

AI motion card: Rewrite the debate motion in your own English and define the three key limits: homework, formative writing and verification.
Vocabulary map: Create a one-page word-family map for six debate terms using only English explanations first.
Evidence labels: Sort the four source cards into experiment, policy guidance and expert argument, then justify each label.
Private voice note: Give a voluntary 60-second spoken claim-evidence-limitation response. Keep the recording local; no upload is required.

Standard

Balanced case file: Build one pro and one con case with equal preparation time and at least two sources each.
Source limitation poster: Design a visual poster showing population, task, tool and outcome for the two experiments. Use your own graphics or openly licensed media.
Rebuttal workshop: Write three fair rebuttals that begin with an accurate summary or concession.
Mini communication round: Run a five-minute monologue plus ten-minute dialogue with a partner and collect criterion-based feedback.

Schwer

Method comparison: Compare the two experimental studies and explain why their different populations and designs matter for transfer to English learning.
Evidence audit: Find one additional teacher-approved peer-reviewed or official source and evaluate its method, relevance, limitations and date. Do not use an AI-generated summary as your evidence.
Policy synthesis: Draft a conditional school guideline that responds to learning, privacy, equity, verification and disclosure without assuming that one study settles the issue.
Abi transfer task: Apply the same claim-evidence-warrant-limitation method to one official 2027 English topic and connect at least three materials independently.

Lernkontrolle

Transfer test: A new study finds that an AI tutor improves vocabulary quiz scores for university students. Explain what would still need to be known before applying the result to Kursstufe writing.
Competing evidence: Two strong studies point in different directions. Develop three plausible reasons for the difference before deciding what follows.
Source-function analysis: Explain how an RCT and a policy guideline can both be useful without serving the same evidential function.
Rebuttal quality: Take a strong argument for the opposite side and write a rebuttal that concedes one valid point, challenges one inference and adds one limitation.
Visual-data transfer: Analyse a new chart by separating description, interpretation and limitation.
Communication transfer: Turn a written paragraph into a five-minute spoken structure with signposting, one question to the partner and one planned concession.

Lernnachweis

For a successful Lernnachweis, demonstrate:

a clearly narrowed debate motion
accurate use of claim, evidence, warrant, counterclaim, rebuttal and limitation
fair representation of competing positions
at least two different evidence types
explicit source checking and transparent citation
differentiated transfer from studies to the exact school context
independent English analysis and writing
appropriate Leistungsfach register and cohesion
visible revision after feedback
privacy-aware and transparent handling of any optional AI support

OERs zum Thema

Sources and further OER-oriented reading

Bastani et al., PNAS 2025: experiment on generative AI and high-school mathematics
Kestin et al., Scientific Reports 2025: randomized study of a designed AI tutor in university physics
UNESCO: human-centred guidance for generative AI in education and research
European Commission: ethical use of AI and data in teaching and learning
European Commission: updated digital-literacy and disinformation guidance
Kultusministerium Baden-Württemberg: official Facherlass Abitur 2027
Landesbildungsserver Baden-Württemberg: communication-exam practice guidance

Folgekurs

Continue with ABI LF - AI source verification lab: trace a claim back to its original source, compare reporting with the study itself, verify quotations and data, and build a compact evidence dossier for a new debate.

Verknüpfte Lernbereiche

ABI LF - AI debate challenge

aiMOOC-Projekte

Frag mich zum aktuellen Lerninhalt oder wähle eine kurze Hilfe.

Textmodus bereit

Sprache