Free Resource
What AI Can and Can't Do
A plain-English breakdown of where AI genuinely helps, where it falls short, and how to use it with confidence.
5 min read
What AI handles well
AI tools are genuinely useful for drafting, brainstorming, summarizing, planning, editing, and repeatable writing tasks. Give the tool a rough set of notes and ask for a first draft. Ask it to summarize a long document into bullet points. Use it to generate ten title options for a blog post. These are tasks where speed and volume matter more than perfect accuracy on the first pass.
Where AI fails
AI is unreliable for anything that requires current information, precise facts, or professional judgment. It will confidently give you a wrong statistic, invent a study that doesn't exist, or miss the nuance that matters most in a legal or medical question. The model doesn't know what it doesn't know. It fills in gaps with plausible-sounding text, which is the specific failure mode you need to watch for.
How to check outputs
Before using any AI output in something that matters, verify the facts independently. If you can't check a claim in under two minutes, rewrite the section without the specific claim. Treat AI output as a first draft from a fast, confident intern: useful as a starting point, not safe to ship without review.
The practical rule
Use AI to reduce the blank-page problem and speed up the first 80% of a task. Spend your own judgment on the last 20%: the accuracy check, the tone adjustment, the decision about what to actually say. That split is the highest-leverage way to use these tools: AI for speed, you for judgment.
Key takeaway
AI is a drafting and brainstorming tool, not a fact source or decision-maker. Use it for speed; apply your own judgment for accuracy.