Bainbridge's Ironies of Automation.
A 40-Year-Old Warning the Church Needs to Hear.

In 1983, a British researcher named Lisanne Bainbridge published a paper in Automatica titled “Ironies of Automation.” It is now the most-cited paper on human-automation interaction ever written — over 4,700 citations and counting.

The paper made a simple, devastating argument: the more you automate a system, the more you degrade the very human skills you need when that automation fails. And automation always fails eventually.

Forty years later, aviation proved her right. Medicine proved her right. Law proved her right. The church is walking directly into the same trap — and almost nobody is talking about it.

Irony One: Humans Are Terrible Monitors

When you automate most of the work, the human's role becomes monitoring the automation. Watching for errors. Reviewing output. Standing by in case something goes wrong.

Bainbridge established a finding that has held for four decades: “It is impossible for even a highly motivated human being to maintain effective visual attention towards a source of information on which very little happens, for more than about half an hour.”

The human brain was not designed for passive monitoring. It was designed for active engagement. When you ask a person to watch a system that works correctly 95% of the time and catch the 5% where it fails — the person will miss it. Not because they are lazy. Because the brain disengages from tasks that require vigilance without active participation.

The ministry application is direct. When a pastor's role shifts from writing the sermon to reviewing the AI's sermon, the pastor becomes a passive monitor. And passive monitors miss errors. Not sometimes. Systematically.

Irony Two: The Skill Degradation Paradox

This is the irony that should keep every church leader awake.

The operator who is supposed to take over when automation fails is the same operator whose skills have degraded from not practicing. Bainbridge stated it plainly: “When manual take-over is needed, there is likely to be something wrong in the process, so the operator needs to be more rather than less skilled than before.”

The person who must intervene when the system breaks is the one least prepared to do so — because the system prevented them from maintaining their skills.

The pattern is universal. Automation improves performance. Dependency on automation degrades the underlying skill. When the automation fails, the operator performs worse than if automation had never been introduced.

Now apply that to a pastor who uses AI for sermon prep every week for two years. The sermons may improve in polish and breadth. But the pastor's ability to exegete a passage independently, to detect doctrinal error without AI assistance, to think theologically from the ground up — that muscle is atrophying. And when the AI produces a subtle heresy — when the “system fails” — the pastor's degraded skill may not catch it.

Irony Three: More Automation Requires More Training, Not Less

The intuitive assumption is that automation reduces training needs. If the machine does the work, the human needs less skill.

Bainbridge demonstrated the opposite. Automated systems require more training investment, not less — because the interventions the human must make are rarer, harder, and more consequential.

In aviation, this has been learned through tragedy. Air France Flight 447 crashed in 2009 because pilots could not manually fly the aircraft when the autopilot disconnected at high altitude. The stall warning activated 75 times. The crew did not recognize a stall. Air France's own internal report identified a “generalized loss of common sense and general flying knowledge” among its pilots.

They had the automation. They had the technology. They did not have the skill to take over when the technology failed. Because the technology had replaced the conditions under which that skill would have been maintained.

“We are creating a world of people who know how to use tools but have forgotten how to think without them.”

The ministry parallel is precise. The more a church uses AI, the more — not less — it should invest in theological education for its leaders. Seminary is not made obsolete by AI. It is made more essential. Because the interventions a pastor must make — catching doctrinal error, discerning when AI output contradicts Scripture, recognizing the impossible backhand of theology — are rarer, harder, and more consequential in an AI-assisted environment.

Irony Four: You Cannot Automate the Human Out

Bainbridge's final irony: automated systems always end up being human-machine systems. You can never fully remove the human. You can only change their role from active operator to passive monitor.

And passive monitoring is the worst use of the human brain.

Research on situation awareness (Endsley & Kiris, 1995) demonstrated that awareness was “significantly lower under fully automated and semi-automated conditions than under manual performance.” The more the system did, the less the human understood about what was happening.

A pastor who writes a sermon knows every contour of the argument. He knows where the text pushed back against his initial reading. He knows which illustration came from a pastoral conversation last Tuesday. He knows what he left out and why. That knowledge — the situational awareness of his own sermon — is lost when AI generates the draft and the pastor merely reviews it.

The Performance-Competence Gap

Horasis (2025) documented what they call the “Performance-Competence Paradox.” People perform better with AI but lose the underlying capability. They look more competent while becoming less skilled.

A pastor using AI produces higher-quality sermons. Performance is up. But the ability to study independently, recognize doctrinal error, and think theologically without assistance is declining. Competence is down. The congregation sees a better sermon. Nobody sees the atrophying theological muscle underneath.

Gartner predicts that by 2030, half of enterprises will face irreversible skill shortages in at least two critical job roles because of unchecked automation. They also predict that by 2026, 50% of organizations will require “AI-free” skills assessments because of critical thinking atrophy from AI dependence.

The secular world is already building guardrails against Bainbridge's ironies. The church has not even started the conversation.

The Pilot's Protocol for Pastors

Aviation solved Bainbridge's paradox with a simple requirement: pilots must hand-fly periodically to maintain proficiency. You cannot rely exclusively on the autopilot. You must keep your manual skills alive.

The church needs the same protocol:

The 40-Year-Old Warning

Bainbridge wrote in 1983. She predicted exactly what happened in aviation, medicine, and law over the next four decades. The pattern is consistent. The evidence is overwhelming. And the church is walking the same path — with 64% of pastors using AI for sermon prep, 73% of churches having no AI policy, and only 12% of pastors feeling comfortable teaching on the subject.

The ironies of automation are not avoidable. They are structural. But they are manageable — if you see them coming. If you build the guardrails. If you maintain the skill.

Bainbridge gave us the warning. Aviation learned the hard way. The church does not have to.

About: AI Fluency Ministry is a project helping the church understand and use AI wisely. OpenLumin is the practical application of that research — a free Bible research companion that retrieves evidence from 15+ scholarly sources so pastors can study with depth and teach with confidence.