Quiz collegato alla guida · Facile Livello

In che modo gli studenti possono verificare i fatti rispondendo al quiz sull'intelligenza artificiale

Verifica la lettura laterale e il metodo SIFT, la triangolazione delle fonti con fonti primarie e il motivo per cui le citazioni e gli autocontrolli dell'IA necessitano di una verifica indipendente.

Domanda 1 di 8

Cos'è la lettura laterale?

Anteprima di ogni domanda8 domande

Le risposte e le spiegazioni vengono visualizzate mentre rispondi al quiz.

  1. Cos'è la lettura laterale?
    • Reading an AI answer slowly from start to finish several times
    • Leaving the source and opening new tabs to see what other trustworthy sources say
    • Reading only the first and last sentences of an answer
    • Asking the chatbot to explain its answer in simpler words
  2. Nel metodo SIFT cosa significa T?
    • Test the chatbot again
    • Translate the claim
    • Trace claims to their origin
    • Trust the most popular source
  3. Perché i chatbot possono produrre citazioni che sembrano reali ma non esistono?
    • They have learned the pattern of a citation and can fill it with invented details
    • They copy citations from a hidden list of fake papers
    • They are designed to test whether students check their work
    • Citations are always pulled from library databases that contain errors
  4. Quale coppia di fonti conta meglio come conferma indipendente di un'affermazione?
    • Two blogs that both quote the same press release
    • The chatbot's answer and the same chatbot's answer to a reworded question
    • A news site and its own sister website
    • A government dataset and a separate university research report
  5. Quali parti di una risposta AI sono a rischio più elevato e meritano di essere controllate per prime?
    • Transition words and general explanations
    • Specific numbers, names, dates, quotes and citations
    • The formatting and headings
    • The opening sentence that restates the question
  6. Why is asking a chatbot 'Are you sure?' not real verification?
    • Chatbots refuse to answer that question
    • It always makes the answer longer
    • The model may just repeat or reverse its answer without checking any outside source
    • It deletes the original answer
  7. What did the 2023 Mata v. Avianca case illustrate?
    • Lawyers were sanctioned for filing a brief with court decisions ChatGPT had invented
    • A court ruled that AI answers count as primary sources
    • A chatbot correctly predicted the outcome of a trial
    • Students may not use AI tools for homework
  8. What is a smart first step to make checking a long AI answer faster?
    • Copy the answer into a second chatbot and trust whichever sounds better
    • Ask the AI to list its factual claims as separate bullet points, then check each one
    • Check only the claims you already believe
    • Rewrite the answer in your own words before checking anything