Segment 1 — The Cold Open

Segment 2 — The Good Intention

Engineers at Airbus and Boeing, along with regulators at the FAA and EASA, introduced full-authority digital engine controls, flight management systems, and envelope-limiting fly-by-wire laws in the 1980s and 1990s. Their explicit goal was to eliminate the most common fatal pilot errors—controlled flight into terrain, loss of control in flight, and unstable approaches—by letting the aircraft itself refuse unsafe trajectories. The historical context was stark: between 1959 and the mid-1980s, human factors still dominated accident statistics even after the introduction of ground-proximity warning systems and better training. Rational actors therefore concluded that further automation would close the remaining gap. They measured the problem in concrete accident categories and saw that adding layers of protection and autoflight could prevent trajectories the pilot alone had not always avoided. No one at the time argued that pilots should stop flying; the premise was that the machine would handle the envelope while the crew retained oversight, a division that appeared both technically feasible and statistically justified by the data then available.

Segment 3 — The Implementation

Airbus launched the A320 family in 1988 with flight-envelope protections that prevented stalls and excessive bank angles in normal law. Boeing followed with its own autoflight suites on the 777 and later types. Early data showed the expected gains: Airbus statistical summaries record an 87 percent drop in controlled-flight-into-terrain fatal accidents once generation-3 aircraft added glass cockpits, terrain awareness, and flight-management computers. Airlines adopted policies that encouraged maximum use of the autopilot to reduce workload and standardize performance. Captain Warren Vanderburgh of American Airlines began warning in 1997 that crews were becoming “children of the magenta line,” but the industry’s priority remained the measurable reduction in accident rates. Training programs therefore emphasized mastery of the flight-management computer and mode awareness rather than routine hand-flying, because the data showed that the automated path produced fewer deviations. The logic was internally consistent: if the protections worked, the safest procedure was to stay inside them.

Segment 4 — The Unintended Consequences

The same systems that prevented most loss-of-control events also removed routine opportunities for pilots to practice manual flight-path management. A 2014 NASA study led by Stephen Casner placed 16 airline pilots in a Boeing 747-400 simulator and found that cognitive skills—tracking position without a map display, integrating raw data, and recognizing mode changes—decayed measurably after periods of heavy automation use, while basic instrument scanning remained relatively intact. The 2009 BEA final report on Air France 447 documented how, after autopilot disconnection, the crew applied sustained nose-up inputs that drove the aircraft into a deep stall; the report noted the absence of high-altitude manual-handling practice as a contributing factor. A later FAA cognitive-skill degradation study (2022–2024) confirmed that declarative knowledge of flight-management and autoflight interactions faded faster than procedural memory when pilots spent twelve to twenty-four months away from the line. Because the envelope protections handled nearly every departure, cruise, and arrival, the average line pilot accumulated far fewer hours actually moving the controls through the full range of normal and abnormal attitudes. When the rare event occurred and the protections stepped aside, the crew therefore faced a task whose cognitive and motor demands had not been rehearsed at the same frequency. Second-order effects appeared in training departments: recurrent simulator sessions focused on automation management rather than raw hand-flying, reinforcing the very dependency they were meant to mitigate. The arithmetic was straightforward: each additional percentage point of flight time spent inside protected modes reduced the sample size of manual recoveries available for skill maintenance, yet the probability of an automation hand-back event, though low, never reached zero.

Segment 5 — The Aftermath

Airbus issued a 2020 manual-flying policy urging pilots to hand-fly segments under safe conditions. The FAA updated Advisory Circular 120-111 to require upset prevention and recovery training, including academic and simulator work on startle effects and energy management. Airlines began inserting deliberate hand-flying exercises and line-oriented flight-training scenarios that force crews to step down automation levels. These measures have not reversed the overall safety gain; Airbus 2025 accident statistics still show generation-4 aircraft maintaining the lowest loss-of-control rates. The correction itself, however, requires ongoing reinforcement, because the underlying automation continues to expand and the daily line environment still favors the protected path. Training departments now treat manual currency as something that must be scheduled and protected rather than assumed to occur naturally through normal operations.

Segment 6 — The Lesson

When Airbus and Boeing chose envelope protection, they rewarded the metric of prevented accidents while leaving the supervisory skill of rapid manual intervention unmeasured and therefore under-practiced. Warren Vanderburgh’s 1997 training directive—that crews must be willing to disconnect at low altitude rather than troubleshoot—remains the operational countermeasure because it directly restores the missing practice loop. Regulators and carriers now schedule recurrent manual-flight segments and upset-recovery drills precisely because the original automation decision had removed those repetitions from daily operations. The same pattern appears today whenever new supervisory systems are introduced: the quantity that improves is the one the machine now owns; the quantity that erodes is the one the human must still be able to seize without notice. What fraction of any new supervisory task will still require the human to intervene at full proficiency, and how will that proficiency be maintained once the system is trusted to handle the routine?

Sources

• ojs.library.okstate.edu — Methods for Preventing the Degradation of Manual Flying Skills in an Automated Cockpit Environment
• rosap.ntl.bts.gov — Cognitive Skill Degradation: Phase III
• aopa.org — Always Learning - Magenta line
• faa.gov — Advisory Circular 120-111 CHG 1, Upset Prevention and Recovery Training
• flightsafety.org — Use It or Lose It
• bea.aero — Safety Investigation into the accident on 1 June 2009 to the Airbus A330-203, flight AF447
• skybrary.aero — Final Report AF447, Airbus A330-203, F-GZCP, Rio de Janeiro - Paris, 1 Jun 2009
• accidentstats.airbus.com — Accident Rates by Category & Generation

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp

Forwarded this email? Subscribe here — it's free.