Separate an unverified claim from a proven falsehood.
Review repeated claims, old notices, synthetic media, traceable citations and local AI-use rules. Match explicit evidence without inferring actual conduct.
Six original scenarios each contribute one match to the authored key. The count describes these supplied facts, local rules and educational concepts, not your actual conduct, character or a certified digital-citizenship level. Every explanation and this full guide is free at every count.
Count evidential origins, not repeated headlines
Five posts can have one underlying source. If they all repeat the same unnamed statement, their number does not supply five independent confirmations. Map where each claim came from before treating repetition as corroboration. A useful new exercise is three messages quoting one leaflet: adding a fourth message does not create another independent observation. This leaves the claim uncertain under the supplied evidence. It does not prove the leaflet false or imply that anonymous sources never report anything accurately.
Preserve the date and the question being checked
A notice can be authentic and still be irrelevant to a claim about today. Compare the visible date, organiser and event with the claim being shared, then look for the current notice through an independently verified source. Editing the old image’s date would create misleading material rather than resolve the question. Conversely, an old date alone does not establish that the original notice was fabricated. Write what has been confirmed now and which part remains uncertain instead of flattening those states into one label.
Convincing sound is not complete provenance
A familiar voice or realistic video does not by itself establish who created a clip, when it was recorded or what was left out. Synthetic imitation is one possibility; missing context in a real recording is another. Seek provenance and relevant independent confirmation before a strong factual claim. Do not replace that process with a promise that a detector will always decide correctly. The practice uses a fictional clip and does not ask you to upload a real person’s voice, face or private recording.
Verify that a source supports the particular claim
A generated citation can look plausible while failing to identify a real supporting source. Check both existence and relevance: finding a real paper with a similar title does not establish that it says what the draft claims. If verification fails, remove the unsupported assertion or state the limitation clearly. Do not invent a page, author or quotation to make the reference appear complete. This is an evidence-handling decision, not a claim that every AI-generated reference or every human-written citation must be false.
Read the local permission and disclosure conditions
The course example permits brainstorming with disclosure and requires the learner’s own checked analysis. All of those conditions matter. Hiding help because the final wording changed would omit the disclosure condition; pasting unverified output would omit the learner’s verification. A different course or workplace may set different rules, so locate the actual current policy before applying the example outside this practice. The reference guidance offers a human-agency context, not a universal permission slip or an official certification of this question bank.
An allowed tool does not authorise every input
A project can permit an AI service for some tasks while withholding approval for confidential transcripts. In that case, payment tier or a changed filename does not override the supplied data rule. Use an authorised alternative or seek the required approval. Do not infer a vendor’s retention, training or contract terms from whether the service costs money. This scenario checks a fictional information boundary; it neither asks for real confidential material nor performs an upload to demonstrate the lesson.
Record what would change your conclusion
For each missed scenario, write the missing evidence or permission rather than a fixed label about yourself. A current original notice, a genuine independent source or explicit approval can change a conclusion; more copies of the same unsupported claim cannot supply those facts. Six matches describe this edition’s authored situations, not media-literacy certification, honesty, intelligence or observed online choices. Keys were author-reviewed against the stated facts and policies. They are revision prompts, without norming, expert-panel consensus or a guarantee that the same short method resolves every real information dispute.