> For the complete documentation index, see [llms.txt](https://theaihandbook.leomohan.net/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://theaihandbook.leomohan.net/chapter-3-where-should-i-not-use-ai.md). # Chapter 3: Where Should I NOT Use AI? ### The “Keep It Simple” Rule **Q1: Should I use AI to calculate 2+2?** **A:** Absolutely not. That’s what your brain (or a basic calculator) is for. Using a massive language model for simple arithmetic is like hiring a jumbo jet to deliver a letter across town—overkill, inefficient, and wasteful. **Q2: Should I use AI to write a one-sentence “Thank You” email?** **A:** No. Just type “Thanks!” It will take you two seconds. Opening an AI tool, crafting a prompt, and copy-pasting the result takes longer than just doing it yourself. It kills the personal touch for no gain. **Q3: Should I use AI for legal or medical advice I’ll act on without a human expert?** **A:** No, and this is critical. AI can hallucinate (make up convincing lies) and has no real understanding of your specific situation. It can summarize information, but it cannot replace a doctor’s diagnosis or a lawyer’s fiduciary duty to you. **Q4: Should I use AI to decide who gets hired or fired?** **A:** No. AI models can inherit human biases from their training data (e.g., penalizing female or minority candidates). Using it as the sole decision-maker is unethical and, in many places, illegal. It should only be used for assistance, not judgment. **Q5: Should I use AI to write a deeply personal message, like a wedding vow or a condolence letter?** **A:** Please don’t. These moments require genuine human emotion and authenticity. An AI-generated message will feel hollow and generic. The recipient wants *your* words, not a machine’s statistically average sentiment. **Q6: Should I use AI to control critical safety systems (like a nuclear reactor or airplane landing) without fail-safe human oversight?** **A:** No. While AI can assist, we don’t yet trust it to handle unpredictable, life-or-death situations with no human in the loop. The consequences of a mistake are too catastrophic. **Q7: Should I use AI for creative work where the goal is raw, unfiltered human expression?** **A:** Probably not. If you’re journaling for therapy, sketching for fun, or improvising on a guitar, the process is the point. AI shortcuts that personal journey and authentic expression. **Q8: Should I use AI to make a decision based on a single, unchanging rule?** **A:** No. If the rule is “If it’s Tuesday, send the report to Bob,” just automate that with a simple script or a calendar reminder. You don’t need a learning system for something that never changes. **Q9: Should I use AI to settle a personal dispute with my partner or friend?** **A:** Absolutely not. AI doesn’t understand the nuances of human relationships, history, or emotion. Outsourcing your argument to a chatbot will likely make things worse, not better. **Q10: Should I use AI when I have zero data to train it on?** **A:** No. AI learns from data. If you have a new, unique problem with no historical examples, AI has nothing to learn from. A human expert using first principles is a much better approach. **Q11: Should I use AI for tasks that require physical dexterity in an unstructured home environment?** **A:** Not yet. While robots are improving, a general-purpose AI that can fold your laundry, unload the dishwasher, and tie your shoes doesn’t exist. It’s far more complex than it looks. **Q12: Should I use AI to replace genuine human connection in customer service?** **A:** Be very careful. Using a chatbot to endlessly deflect a frustrated customer who just wants to talk to a human is a recipe for disaster. AI should *assist* human agents, not build a wall around them. **Q13: Should I use AI to predict the stock market with 100% accuracy?** **A:** No, and anyone selling you this is lying. If a reliable AI could do this, its creator would keep it secret and become the richest person on earth, not sell it to you. Markets are chaotic and influenced by human emotion. **Q14: Should I use AI for a task that is perfectly fast and efficient already?** **A:** No. Don’t fix what isn’t broken. If your current process works great, introducing AI just adds complexity, cost, and a new potential point of failure. **Q15: Should I use AI to generate news articles without fact-checking?** **A:** No. This is how misinformation spreads. AI doesn’t know what’s true; it knows what’s statistically plausible. It needs a human editor to verify every fact before publication. **Q16: Should I use AI to write my child’s college admission essay?** **A:** No. This is ethically dubious and easily detectable. The essay is meant to reflect the student’s unique voice and experiences. Using AI is misrepresentation and could ruin their chances. **Q17: Should I use AI when I need complete and perfect accuracy?** **A:** No. AI is probabilistic, not deterministic. It gives you a “best guess.” For tasks like accounting, where numbers must balance perfectly, a rule-based system is still superior. **Q18: Should I use AI to automate a process I don’t fully understand myself?** **A:** No. If you don’t understand the steps, you won’t know when the AI makes a mistake or goes off track. You become a helpless supervisor of a system you can’t control. **Q19: Should I use AI to replace thinking in school (e.g., doing your math homework for you)?** **A:** No. The purpose of homework is for *you* to learn and practice. Using AI to do the work for you is cheating yourself out of an education. It’s the tool doing the learning, not you. **Q20: Should I use AI to generate a recipe using every ingredient in my fridge at once?** **A:** No. This is a classic example of AI missing common sense. It might suggest “chocolate-covered sardine ice cream” because technically, those are all ingredients. AI has no concept of what tastes good together. *** 💬 Enjoyed this chapter? Have questions or thoughts?\ Join the discussion on GitHub → [**Click here to Comment**](https://github.com/leomohan/theAIhandbook/discussions) --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://theaihandbook.leomohan.net/chapter-3-where-should-i-not-use-ai.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.