< Back

Bot Don’t Kill My Vibe

Competencies: Critical Analysis, Fluency, and Evaluation; Skilling and Productivity

Overview

In this lab, you’ll learn to use AI systems as a constructive critic to identify potential problems, weaknesses, or overlooked issues in your work. By having AI tools take an adversarial stance, you’ll practice extracting valuable feedback while learning to evaluate and filter AI critiques.

Details

Choose something you’ve created or are working on where you’d genuinely value critical feedback. This could be writing, an idea, or a product.

Use a chat bot to take on an adversarial role, finding potential problems or weaknesses in your work. The goal isn’t to discourage you, but to help strengthen your work by identifying areas for improvement before they become issues.

You’re using it to preempt criticism - what problems could others find?

The ultimate way you adapt to criticism is up to you - this just gives you alternate angles to think through your work. Sometimes, this leads to useful changes. Other times, it strengthens your resolve. In writing, it may not be a change that you implement, but something you address and justify.

Some examples - you don’t have to follow any of these:

  • Criticizing the structure, flow, or argument of an essay
  • Noting problems with tone in an email
  • Finding potential problems or vulnerabilities in a card game that you’re designed (credit Emily G. for the example)
  • A plan for a life goal or endeavor

Example Prompts - again, no need to use these specific ones:

  • “What could go wrong with this approach?”
  • “What critical assumptions am I making?”
  • “Take a skeptical view of this and explain why it might fail”
  • “What obvious things might I be missing?”
  • “How could this be misinterpreted or misused?”
  • “From the perspective of a hostile reddit-er, critique this”

Feel free to develop your own approaches to soliciting critical feedback. Be creative in how you frame the critique – you might have the AI take on specific personas or roles relevant to your work.

The goal isn’t just to receive criticism, but to learn how to effectively solicit and filter constructive feedback using AI tools as a supplement to human review, or as a way of interacting with out your own internal thoughts and consideration.

Some tips:

  1. Start Broad, Then Focus: Begin with general criticism, then drill down into specific concerns that emerge.
  2. Push Back: If the AI’s criticism seems shallow or unhelpful, ask it to elaborate or approach the problem from a different angle. Or go back and restart the chat earlier.
  3. Filter Feedback: Not all criticism will be valuable. Practice distinguishing between helpful insights and less useful feedback.
  4. Iterate: Consider using the feedback to make improvements, then submit the revised version for another round of critique. (updates don’t need to be made for submission)

Completion Requirements

Try this exercise with multiple initial prompts.

You don’t need to update your idea/text/code/product in response to the critique. If it’s not sensitive or too vulnerable, you can share snippets of the chat and a link, but it’s not necessary.

To complete this lab, submit:

  1. A brief description of what you analyzed, and what type of prompt worked best (i.e. the wording)
  2. 2-3 key pieces of valuable feedback you received, and optionally unhelpful feedback. You can summarize.
  3. Short reflection on which feedback was helpful vs unhelpful. Only one or two sentences needed - completing the exercise is more important.

Submit these items on Canvas in the discussion post.

Portfolio Requirements

Size: 1 point

If you choose to include this lab in your portfolio, submit the same information as above, plus one more point:

  1. A comment on which prompted approaches yielded the most feedback (1-3 sentences).

Grading will focus on: quality of analysis and reflection in your lab submission; depth of critical thinking in analyzing AI feedback; clear articulation of which feedback approaches were most/least effective; thoughtful reflection on the overall experience of using AI as a critic.