The Better the Autopilot, the Worse the Pilot — Julien Reszka
← Blog
Automation doesn't make operators more careful. It makes them forget how to be. The more reliable the system, the less ready the human.
80% Aviation accidents attributable to human factors Boeing Statistical Summary
The argument for automation is that it frees up cognitive bandwidth. Fewer routine decisions means more headroom to think carefully about the ones that matter. What actually happens is the opposite: when a system reliably handles a task, the human monitoring it gradually stops monitoring, because nothing ever goes wrong, and sustained attention without feedback is not something brains do voluntarily. Aviation has a name for this: automation-induced complacency, and it shows up in accident reports where pilots failed to notice system failures they would have caught immediately if they had been flying manually. The irony is that the better the automation, the worse the problem: a system that almost never fails produces operators who are almost never ready for the moment it does. The countermeasure is deliberate:
Identify the critical tasks you have handed to automation.
Periodically turn them off and practice manually.
Keep the interval short enough that the skill does not decay.
Not because the machine is unreliable, but because you are, and the only way to stay ready is to keep practicing the skill the machine has been covering for you.
Myth: More reliable automation makes human operators saferParasuraman & Manzey, Human Factors, 2010
Schedule regular manual practice for any critical task you have automated. The interval should be short enough that the skill does not decay before the next failure.
Post on X
Discussion
Are you or your team losing the manual skills and situational awareness that your automated systems cover for you?
Post on X
Patrick D. London, UK 2026-04-14
Our on-call engineers haven't had a real incident in eight months. I can feel the readiness eroding. Scheduling a game day this week.
Reply
Ava S. Amsterdam, Netherlands 2026-04-15
We run quarterly chaos drills for exactly this. The skill comes back quickly but you have to keep the interval short or it's back to square one.
Reply
Yara M. Austin, TX 2026-04-16
Hadn't framed it this way before but it's accurate. Haven't touched a manual deploy in over a year and I know I'd be slow and second-guessing everything if CI broke tomorrow. Trying to pick one task a week to do by hand.
Reply
Daniel O. Singapore 2026-04-18
The aviation parallel is apt but I'd push on one thing: pilots are required by regulation to log manual hours. Most engineering orgs have no equivalent obligation so the decay is invisible until something breaks.
Reply
Tomás V. Barcelona, Spain 2026-05-09
Coming to this late, found it via the automation series. We ran a game day last week for the first time in eighteen months and the results were uncomfortable. Two engineers couldn't complete rollbacks they'd have done in their sleep three years ago. Scheduling quarterly from here.
Reply
Your name *
Your email * (never published)
City, country (optional)
Tell me about your situation *
Send<br>All comments are manually moderated by the author.
Subscribe to get new posts by email →
See also
The Future of Work Is Getting Out of the Way
Bend Under Pressure, Not Under Nonsense
Stop Pasting Files Into Agent Prompts
Browse all figures →<br>Browse all myths →