Every AI reflects training-data biases. Check: gender, culture, economic class.

Set clear policy: which decisions will never be delegated to AI in your team.

Keep drawing as joy not test — don't say "that's pretty" or "what's that?" — ask about the story.

Before relying on AI in specialist field, show to human expert — 30% of time uncovers core error.

Every factual AI claim = ask for a source. No source = don't use.

When delegating a multi-step chained task to the model, define mandatory human-review checkpoints before continuing.