U1_Lesson4_Bias In The Machine.

Purpose: As students develop a 'Recipe Bot', they experience how bias can implicitly and unintentionally creep into large-language models.

Vocabulary: Bias

Journal: What is Bias? State where you have either experienced or demonstrated bias?

Activity: Log onto Code.org: U1_Lesson4_Bias In The Machine.
Goal: I You haven hired to help create a brand new language model called RecipeBot. This chatbot will help people discover new recipes or suggest different types of cuisines based on people's preferences. Our job is to curate the online data sources used to train the model. This means looking at a lot of food websites!

• 1. Go online, find a recipe website, and then find an article posted on the website.
  • Example: "I offered the first article that I found: allrecipes.com/Meet the Women Who Figured Out What 63 National Parks Taste Like/Nina Elder and Melissa Knific/February 8, 2026/Summary: Recipes reflecting the regions of 63 National Parks."
• 2. Identify at least 10 recipe-related words that a language model could learn from this article.
  • Example: cookbook, ingredients. dishes, food, sandwich, grill, slow cooker, flavor, bake, scent, taste.
• 3. Our new job is to identify potential sources of bias in these resources. G
  • Google "Man vs. Food." Adam Richman travels to different parts of the US where he psyche himself up to eat usually a large amount of food and challenging like a burger with super hot chilis.
  • Answer the online questions regarding Bias and if this site should be offered to our Bot as training data?
• 4. UNESCO?
  • Is there bias? Should we offer this site to our Bot as training data?
• 5. Ina Garten?
  • Is there bias? Should we offer this site to our Bot as training data?
• 6. Question: Is it more valuable to include Everything found on the internet about recipes?

Video: AI and Bias.