Perfilado de sección

  • LESSON 8 - ARTIFICIAL INTELLIGENCE FOUNDATIONS FOR BROADENING CULTURAL HORIZONS

    What are stereotypes in AI models?


    In this lesson, we address stereotypes, also known as social biases, in language models. What is a stereotype? They are generalizations about groups of people often built from limited information or prejudices. If AI models learn from data produced by our societies, isn't it expected that they also learn our prejudices? We analyze how biases originate in training data (which is often non-transparent) and how the strong dominance of English and European languages excludes a large portion of the world's linguistic and cultural diversity.

    We also distinguish between in-group and out-group stereotypes, observing concrete examples of how these can appear in automated responses. The goal is to recognize that AI is not neutral and that its outputs can reinforce existing inequalities. Through a practical activity, we propose identifying in-group stereotypes and reflecting on out-group stereotypes to understand how technology that relies on stereotypes can impact the educational practices we build.

    At the end of the lesson, you will be asked to answer a few questions. 

    👩‍🏫 Instructor: Luciana Benotti

    • What are stereotypes in AI models? URL

      Prefer to watch in a language other than Spanish? Just turn on CC, go to Settings (⚙️) > Subtitles > Auto-translate, and select your language!

    • Below you will find the lesson slides, designed as a reference to revisit the ideas, questions, and tensions raised in the video. (Please note: The slides are in Spanish).

      📚 Key Concepts

      Classification Error, False Negative: When the system outputs "NO" (e.g., "This was written by a human") but the real answer is "YES" (it was generated by AI).

      Classification Error, False Positive:
      When the system outputs "YES" (e.g., "This text was generated by AI") but the real answer is "NO" (it was written by a human, such as the Argentine Constitution).

      Retrieval-Augmented Generation (RAG):
      A technique that allows a language model to query trusted external sources (such as a specific PDF or a medical database) before generating a response, reducing (though not eliminating) hallucinations.

      Prompt Instability:
      A phenomenon where tiny changes in the prompt (a comma, a synonym) drastically alter the output. The script emphasizes that this is a limitation in language model design rather than a user skill deficiency.

      Context Window:
      The maximum limit of information a model can retain in its memory during an interaction. If a text exceeds this limit, the model "forgets" part of it.

      A creation by Fundación Vía Libre in collaboration with FAMAF – UNC.
    • Assignment: What are stereotypes in AI models? Cuestionario