Automation Bias

We give excessive trust to automated recommendations and may overlook contradictory evidence.

Automated systems can appear consistent, complex and authoritative. Reliance becomes more likely when people are busy, inexperienced, under pressure or accustomed to the system being correct. Note that automated tools, including AI, can make mistakes.

• A driver follows navigation into a restricted road despite visible signs.
• A reviewer accepts an AI-generated summary without checking the original document.
• A professional follows a software alert while overlooking conflicting information on the screen.

• Verify high-impact outputs independently.
• Check the original source.
• Ask what data the system may be missing.
• Watch for unusually confident answers.
• Maintain a clear human-override process.