Lesson Content

Six core AI capability categories.

1. Language generation and editing AI excels at producing well-written text to specification: drafts, summaries, rewrites, format conversions, tone adjustments. The underlying language pattern learning is directly applicable here. This is AI's strongest consistent capability.

2. Information synthesis and summarization AI can take large amounts of text – documents, meeting notes, research papers – and produce structured summaries, key point extractions, and cross-document comparisons. The ability to process long context and identify patterns makes this reliable.

3. Explanation and translation AI explains complex topics in simpler terms, and simplifies specialist language for general audiences. It also translates across languages at a quality level suitable for many professional contexts (though high-stakes translation still benefits from professional human review).

4. Code generation and review For common programming languages and well-documented frameworks, AI generates working code, identifies bugs, explains code function, and translates between languages. Coverage varies by language – common languages (Python, JavaScript, SQL) perform better than obscure ones.

5. Structured analysis and frameworks When given clear criteria and scope, AI applies structured analytical frameworks: comparing options, evaluating against criteria, generating pros/cons analysis, producing decision matrices. The key is providing the criteria – AI is good at applying frameworks, not always at selecting the right one.

6. Brainstorming and ideation AI generates options, alternatives, and creative angles quickly. As a brainstorming partner, it produces a breadth of possibilities for human judgment to evaluate and select from. It is better at generating options than evaluating which one is best.

What makes each category genuinely strong.

All six categories share a common characteristic: they are language-intensive tasks where high-quality examples were well-represented in training data. Writing, synthesis, explanation, code, analysis frameworks, and brainstorming – all of these have extensive training coverage. This is not coincidence; it is the pattern that defines genuine AI capability.

Common overpromises to recognize.

AI marketing frequently overstates: autonomous decision-making capability, reliable factual research without verification, emotional intelligence equivalent to human empathy, strategic judgment, creativity that matches human originality. These are areas where AI has surface-level performance that does not hold up in high-stakes, nuanced application.