Learning objectives

• Apply the Pause – Try – Check – Ask routine.
• Identify red-flag instructions in AI answers.
• Use official help pages and trusted people to verify sensitive tasks.
• Build recovery habits when instructions do not match the screen.

AI instructions should be followed slowly. A safe routine for learners is Pause – Try – Check – Ask. Pause before starting. Try only one small step. Check whether the screen matches the instruction. Ask for help if the instruction involves money, health, passwords, official accounts or personal information.

Some AI instructions should raise immediate caution. Learners should stop if the AI suggests ignoring a warning message, turning off security settings without understanding why, sharing a one-time code, sending money, installing an unknown app, or entering personal details into an unfamiliar page. In these cases, official support and trusted people are more appropriate than AI advice.

Educators should practise “mismatch moments”. This means intentionally showing learners that the AI’s step does not match the screen. The trainer then models a calm response: “The instruction does not match. I will not panic. I will describe what I see and ask the AI to adjust, or I will check the official help page.” This turns errors into learning opportunities.