KoreField
AI Foundation School/Phase 2: Quick Wins & Confidence

Module 5: Thinking With AI (Fixing Outputs)

1 dayTeach debugging mindset

AI Avatar Lesson

Video will be available when Cloudflare Stream is configured

1 day
Coming Soon

AI Gets Things Wrong — And That's Normal

AI models produce incorrect, incomplete, or misleading outputs regularly. This is not a bug — it is a fundamental characteristic of how language models work. They generate text based on statistical patterns, not factual verification. The skill that separates effective AI users from everyone else is the ability to identify problems in AI output and fix them systematically.

Common AI Output Problems

  • Hallucination: confident statements that are factually wrong
  • Vagueness: generic responses that lack specific, actionable detail
  • Wrong format: output that does not match what you asked for
  • Missing context: AI ignores important constraints from your prompt
  • Outdated information: AI training data has a cutoff date

The Fix-It Framework

When AI output is wrong, do not start over. Debug it. Step 1: Identify what is wrong (factual error, wrong format, missing detail). Step 2: Determine why (vague prompt, missing context, task too complex). Step 3: Fix the prompt (add specificity, provide examples, break into smaller steps). Step 4: Regenerate and compare. This iterative approach is faster and more effective than rewriting from scratch.

Debugging AI output is the same mindset as debugging code. Identify the symptom, trace the cause, apply the fix, and verify the result. This is a core professional skill for anyone working with AI.

Key Takeaway

AI outputs are often imperfect. The professional skill is identifying what went wrong and fixing it systematically — not starting over. Use the Fix-It Framework: identify, diagnose, fix the prompt, regenerate, and compare.

Review Questions

1. What is the first step in the Fix-It Framework for AI outputs?

2. What is a hallucination in AI output?