Modern transport-category aircraft arrive at the flightdeck loaded with technology: Primary Flight Displays (PFDs), Multi-Function Displays (MFDs), Flight Management Systems (FMS), autopilots capable of flying from shortly after takeoff to touchdown, and autothrottle systems that manage thrust across every phase of flight. This suite of automation has unquestionably improved safety margins and reduced crew workload during routine operations. Yet it has simultaneously introduced two closely related, genuinely dangerous phenomena: automation bias and automation complacency. Understanding them in depth—their definitions, their mechanisms, their operational consequences, and the CRM strategies that counter them—is essential knowledge for every Airline Transport Pilot certificate candidate and for every working crew member.
Defining the Hazards
Automation bias is the tendency to over-trust automated systems, accepting their outputs without adequate independent verification. A crew experiencing automation bias treats the FMS route, the autopilot-commanded pitch, or the autothrottle-selected thrust as inherently correct and reduces or eliminates the cross-checking that would catch an error. The FAA's Risk Management Handbook (FAA-H-8083-2) identifies this as a form of channelized attention—mental focus so concentrated on trusting the machine that contrary cues are filtered out or ignored.
Automation complacency is related but distinct. Where bias is an attitude toward the system's outputs, complacency is an erosion of active, vigilant monitoring over time. Because modern autopilots perform correctly on the overwhelming majority of flights, the brain's threat-detection circuitry gradually down-regulates. Monitoring becomes passive rather than active. The crew watches the displays without truly processing what they show. The Pilot's Handbook of Aeronautical Knowledge (FAA-H-8083-25) warns that glass cockpit pilots must continuously validate what the automation is doing, understand why it is doing it, and remain prepared to assume immediate manual control—precisely because passive observation is insufficient.
How Automation Bias and Complacency Develop
The reinforcement cycle
Every uneventful flight reinforces the unconscious belief that the automation will continue to perform flawlessly. This is not irrationality—statistically, it usually does. But the reinforcement cycle progressively raises the threshold at which a pilot will question or override the system. The result is a crew that responds more slowly to anomalies and may fail to recognize them at all until they have cascaded into a serious deviation.
Out-of-the-loop syndrome
When automation handles most flying tasks, the crew's situational awareness can degrade because they are no longer actively building and updating a mental model of the flight. This is called being out of the loop. FAA-H-8083-2 describes how pilots who are out of the loop take longer to detect problems and are slower to formulate correct responses when the automation fails or transitions to an unexpected state. The insidious aspect is that a crew can feel highly aware—the displays are beautiful and information-rich—while their cognitive engagement with the actual flight path, energy state, and system status is superficial.
Mode confusion and mode surprise
Glass cockpit autopilots operate across many distinct modes: LNAV, VNAV, altitude select, altitude hold, approach mode, go-around mode, speed protection modes, and more. Transitions between modes can be triggered automatically by events the crew did not consciously initiate—an altitude capture, a glideslope intercept, a speed limit in the FMS, or a temporary exceedance. When a crew is not actively monitoring the Flight Mode Annunciator (FMA), an unexpected mode change—commonly called a mode surprise—can go unnoticed. The aircraft may be climbing when the crew believes it is leveling, or decelerating below a target speed, with neither pilot aware because neither is actively cross-checking.
Why It Matters: Operational and Legal Consequences
Automation-related errors have appeared in the causal chains of numerous transport accidents. A crew distracted by FMS reprogramming may miss a subtle pitch deviation that, over several minutes, produces a significant altitude bust. An incorrect altimeter setting entered without cross-check can cause the autopilot to level off hundreds of feet from the cleared altitude. In reduced-visibility instrument conditions, these errors compress the margin for recovery to near zero.
From a regulatory standpoint, no degree of automation sophistication transfers any responsibility away from the crew. 14 CFR 91.3 makes the Pilot in Command solely responsible for the safe conduct of the flight. 14 CFR 91.13 prohibits careless or reckless operation. An autopilot malfunction is not a legal defense for a controlled flight into terrain. The PIC is always accountable, which means active management of the automation is not optional—it is a legal obligation.
CRM Strategies That Counter These Hazards
Mode awareness as a habit
Effective CRM requires that every mode change be verbalized and confirmed by both pilots. After any FMS input, autopilot engagement or disengagement, or altitude change clearance, the monitoring pilot should read back what the FMA displays and confirm the armed and active modes match the crew's intentions. This two-pilot cross-check interrupts the passive-monitoring pattern before it takes hold.
Structured automation levels
FAA-H-8083-2 and CRM doctrine both emphasize that crews must consciously decide which level of automation to use for any given phase of flight rather than defaulting to maximum automation. Using a lower automation level—hand-flying with raw data, or flying with autopilot but without FMS guidance—during lower-workload phases keeps the crew mentally engaged and preserves the cognitive loop that detects anomalies. Deliberately hand-flying occasional approaches or segments in cruise, when conditions and SOPs permit, maintains manual proficiency and cognitive engagement simultaneously.
Challenging and cross-checking culture
CRM teaches that either crew member must feel empowered to verbalize concern when something does not look right. A first officer who notices the autothrottle is not advancing as expected must speak up, and the captain must receive that input without defensiveness. CRM training guidance in AC 120-51 and the Risk Management Handbook (FAA-H-8083-2) underscores that a culture of open challenge and verification—rather than deference to automation or to authority—is the primary defense against errors that automation makes invisible.
Monitoring discipline and sterile cockpit integration
During critical phases of flight—including all ground operations, taxi, takeoff, landing, and other flight operations conducted below 10,000 feet—14 CFR 121.542 (the sterile cockpit rule) prohibits non-essential activities that distract from flying duties. This regulation directly addresses the attention management problem. At higher altitudes, crews must create their own discipline. Dividing monitoring responsibilities explicitly—one pilot heads-down for FMS programming while the other maintains positive outside and instrument scan—prevents both crew members from being simultaneously out of the loop.
Key Numbers and Rules
- 14 CFR 91.3: PIC is solely responsible for safe conduct of the flight regardless of automation state.
- 14 CFR 121.542: Sterile cockpit rule prohibits non-essential activities during critical phases of flight, including ground operations, taxi, takeoff, landing, and other flight operations below 10,000 feet MSL during Part 121 operations.
- FMA scan: Both pilots should verify armed and active modes after every autopilot or FMS input—this is not optional CRM etiquette; it is the primary defense against mode surprise.
- Out-of-the-loop risk: FAA-H-8083-2 identifies that pilots who have been out of the loop take significantly longer to recognize and respond to automation failures or unexpected mode transitions.
- Complacency as a risk factor: FAA-H-8083-2 discusses complacency as a distinct risk factor tied to overconfidence and reduced vigilance—it is not one of the FAA's five formally named hazardous attitudes—and the recognized antidote is active vigilance and disciplined monitoring, not simply reducing workload further.
Common Test Traps
- Automation does not equal reduced PIC responsibility. The ATP knowledge test routinely tests the misconception that engaging the autopilot or FMS delegates legal responsibility. It does not. 14 CFR 91.3 is unambiguous.
- Complacency is a distinct risk factor, not just fatigue. Questions may present a scenario where a well-rested, low-workload crew fails to detect an error. The cause is complacency—overconfidence and reduced vigilance with a specific antidote—not physiological fatigue.
- Mode awareness is a CRM skill, not just a technical skill. Scenarios involving unexpected mode transitions test whether you understand that the correct response is a coordinated two-pilot verification of the FMA, not just a solo pilot correction.
- Reducing automation can increase safety. Test items may imply that using more automation is always safer. The correct FAA position is that crews should select the appropriate level of automation for the situation, and that sometimes a lower level is the correct choice to maintain engagement and proficiency.
- Manual flying proficiency is a safety of flight issue. FAA-H-8083-3 links over-reliance on automation with degraded hand-flying skill. Questions associate infrequent manual flying with elevated risk during automation failures, go-arounds, and emergencies requiring immediate manual takeover.
Memory Aid
Many line crews use the three-question discipline sometimes described in CRM training as "What, Why, Want": after every automation input or mode change, each pilot should quickly confirm—What is the automation doing right now? Why is it doing that? Is that what we want? This loop transforms passive observation into active management and is directly consistent with the monitoring philosophy promoted throughout FAA-H-8083-2 and FAA-H-8083-25.