Working in AI · Contents

Understand AI

Knowing what AI is good and bad at

Before you hand anything over: which tasks suit AI, where it tends to fail (made-up facts, stale knowledge, confident wrong sums), how much effort to put into asking, and which tool fits the job.

Is this a task AI is suited to, and which tool?

Its four sub-skills

• Map AI's strengths: What AI is genuinely good at, and why.
• Map AI's failures: Where it goes wrong: made-up facts, bias, its limits.
• Manage AI's effort: Giving it the context, breaking the work down, iterating.
• Match AI tool: Choosing the right tool or model for the task, or none.