What AI Can — and Cannot — Do
Use AI faster by knowing which jobs fit it well — and where human verification still matters.
Compress long material into a useful overview.
Rewrite, translate, restructure and reformat information.
Create drafts, ideas, code, images or structured outputs.
Group, label and extract patterns from inputs.
The useful mental model
Modern generative AI is powerful at producing plausible outputs from patterns in its training and context. That makes it useful for drafting, transforming and exploring information — but a fluent answer is not automatically a verified answer.
Draft and transform
Summaries, first drafts, brainstorming, formatting, code scaffolding and extracting structure from messy text.
Facts with consequences
Current facts, legal or medical decisions, exact citations, financial claims, or anything where a confident mistake would matter.
AI can sound certain and still be wrong
One of the most important beginner skills is separating fluency from reliability. If the answer depends on fresh facts, exact numbers or authoritative sources, verify it against the relevant source.
Low-risk use: “Turn these meeting notes into five action items.”
Higher-risk use: “Tell me the exact legal filing deadline for my case.”
The second task may still benefit from AI for organizing questions, but the final answer should come from an authoritative source.
A simple decision rule
Use AI aggressively for drafting and transformation. Use verification aggressively when facts, freshness or consequences matter.