Cuestionario vinculado a una guía · Medio Nivel

Abogados sancionados por un cuestionario sobre citaciones de casos elaborado por IA

Prueba los hechos de Mata v. Avianca, por qué los modelos lingüísticos inventan citas y qué hábitos de verificación detectan autoridad falsa antes de presentar la solicitud.

Rutas de guía relacionadasAbogados sancionados por citaciones de casos falsos de IA
Pregunta 1 de 8

En Mata v. Avianca, ¿qué pasó cuando el abogado investigador preguntó a ChatGPT si los casos citados eran reales?

Vista previa de cada pregunta8 preguntas

Las respuestas y explicaciones aparecen a medida que realiza el cuestionario.

  1. En Mata v. Avianca, ¿qué pasó cuando el abogado investigador preguntó a ChatGPT si los casos citados eran reales?
    • ChatGPT said the cases were real
    • ChatGPT refused to answer
    • ChatGPT flagged the cases as fabricated
    • ChatGPT directed him to Westlaw
  2. ¿Qué sanción impuso el juez Castel en Mata vs. Avianca del 22 de junio de 2023?
    • Disbarment of both attorneys
    • A $50,000 fine against the client
    • Dismissal of the lawyers from the bar association
    • A $5,000 penalty on the two lawyers and their firm, plus letters to the client and falsely named judges
  3. Según la guía, ¿qué identificó el juzgado de Mata como el verdadero fracaso de los abogados?
    • Using any AI tool for legal research
    • Abandoning their responsibilities and standing by the fake cases after being questioned
    • Filing the case in the wrong district
    • Missing a filing deadline
  4. What lesson does the guide draw from Wadsworth v. Walmart?
    • Only self-represented litigants file fake citations
    • Judges no longer sanction AI errors
    • Citations from a firm's internal AI tool can also be fabricated
    • Google Bard is more reliable than ChatGPT
  5. Why can a language model produce a fake citation that looks convincing?
    • Legal citations follow a highly regular format the model can imitate
    • Models copy citations directly from court dockets
    • Courts publish fake citations for training
    • Citations are too short for models to get wrong
  6. Why does asking the same model to confirm its own citation fail as a check?
    • The model deletes its earlier answer
    • Models cannot read their own output
    • Confirmation questions cost extra tokens
    • It runs the same generation process that produced the error
  7. Which step counts as real verification of an AI-supplied case citation?
    • Asking a second chatbot whether the case exists
    • Checking that the citation format looks correct
    • Pulling the opinion from an authoritative source, reading the passage and running a citator
    • Confirming the case name appears in the AI's summary twice
  8. In Park v. Kim (2024), what did the Second Circuit do about a brief citing a nonexistent decision?
    • Awarded the opposing party attorney fees of $1 million
    • Referred the attorney to its grievance panel
    • Ordered the case retried before a new judge
    • Banned AI use by all lawyers in the circuit