There is a version of automation use that looks like good airmanship and isn't. It is the crew that engages the highest available level of automation at every opportunity — LNAV, VNAV, autothrust, the full stack — and then monitors it with varying degrees of attention. The logic, if it is ever made explicit, is that more automation equals less workload. Engage everything, step back, let the systems fly the aircraft.
The problem with that logic is that it is sometimes right and sometimes catastrophically wrong. And the difference between those two cases is precisely what selects appropriate level of automation is about. It is not a behaviour about using more automation or less automation. It is a behaviour about asking a specific question before every mode selection: will this reduce workload in the conditions I am actually in?
The Automation Ladder
Modern transport aircraft offer a spectrum of automation levels, and understanding that spectrum is the prerequisite for selecting intelligently within it. At the top end, full flight management system engagement — LNAV tracking a programmed lateral path, VNAV managing the vertical profile, autothrust maintaining target speed — the aircraft is flying the plan. The crew's role is to monitor, manage, and intervene when required.
Step down the ladder and the picture changes. Selected heading mode rather than LNAV. Selected altitude rather than VNAV. Speed selected rather than managed. Each step down transfers a degree of path management back to the crew — not to their hands necessarily, but to their active inputs and judgement. Step further and you remove autothrust, which changes the energy management picture entirely. The aircraft is still automated, but the automation is responding to crew selections rather than executing a pre-planned profile.
None of these levels is inherently superior. Each is appropriate in some conditions and less appropriate in others. The crew that understands the ladder — what each level does, what it demands of the crew in return, and where its limits are — is the crew that can select intelligently. The crew that defaults to the top of the ladder regardless of conditions has stopped selecting at all.
More automation does not always mean less workload. Sometimes it means a different kind of workload — and in dynamic conditions, that difference matters enormously.
When Higher Automation Increases Workload
The conditions under which full automation management genuinely reduces crew workload are well understood: a stable, programmed flight profile, a cooperative ATC environment, and a situation that is developing slowly enough for the automation to track it effectively. Cruise, a planned approach in good conditions, a departure following the filed SID. In these conditions the automation does what it was designed to do and the crew benefits accordingly.
The conditions under which full automation can increase workload are less often discussed. When the situation is sufficiently dynamic — rapid ATC re-clearances, vectors that interrupt the programmed profile, turbulence that degrades autothrust effectiveness — the automation may not be able to keep up. The crew finds themselves making frequent FMC inputs to chase a changing clearance, watching the autothrust hunt in turbulence, or monitoring a VNAV profile that has become disconnected from the actual requirement. The cognitive load of managing the automation in these conditions can exceed the cognitive load of flying a selected mode manually and directly.
Autothrust in significant turbulence is a practical example. On a smooth day, autothrust maintains speed precisely and removes a monitoring task from the crew. On a turbulent day, the autothrust may respond to airspeed fluctuations with continuous thrust lever movement — adding noise to the power picture, occasionally chasing transient speed deviations in ways that feel counterproductive. Some crews in some conditions find it cleaner to disconnect autothrust and manage thrust manually, accepting a direct workload input in exchange for a more predictable and controllable energy picture. That is not a failure of airmanship. It is exactly the judgement the behaviour requires — but it carries a caveat.
Choosing a lower automation level is only a valid option if you have the currency to execute it well. Disconnecting autothrust in turbulence is a reasonable choice for a crew that manages manual thrust regularly and confidently. For a crew that rarely flies without autothrust, the same choice may introduce more workload than it removes — and may produce worse energy management, not better.
This is the dependency that automation creates over time. The automation that makes the routine operation efficient can, if it becomes the only way of operating, quietly erode the capability needed to operate without it. Selecting appropriate automation level requires honest self-assessment of what your actual capability is at lower levels — not what it was at type rating, but what it is now.
The Tendency to Remove Automation at the Wrong Moment
There is a specific failure mode in automation management that is worth naming directly. It is the tendency to disconnect automation at precisely the moment it is most useful — because it appears to have failed to do the job.
An overspeed trend developing on the approach. An underspeed during a demanding departure. The autothrust appearing to lag behind the required response. The instinct, and it is a strong one, is to take direct control — to disconnect and fly manually because the automation is not doing what you need. That instinct is sometimes right. It is also sometimes the fastest route to making the situation worse.
In the overspeed and underspeed cases on a modern aircraft, the autothrust will in the majority of circumstances execute the correct recovery if given the time to do so. The crew member who disconnects and applies manual thrust at that moment is taking on a task that the system was already handling — and is now managing both the recovery input and the consequences of having disconnected the automation, under exactly the workload conditions that prompted the disconnection in the first place.
The question to ask is not "is the automation doing what I expect right now?" It is "will the automation resolve this, and will it do so more effectively than I can manually in these conditions?" That is a harder question, and it requires knowledge of what the automation actually does in that situation — which takes us back to the currency and knowledge requirement that underpins the whole behaviour.
The Exception That Changes Everything
Everything written above argues for caution about removing automation under pressure. There is one scenario where the opposite is true, and it needs to be stated clearly.
Jet upset and unusual attitude recovery — the domain of upset prevention and recovery training — is the scenario where the instinct to retain automation is precisely wrong. In a developed unusual attitude, particularly at high altitude where the aerodynamic margins are narrow and the consequences of incorrect inputs are severe, the correct response is to disconnect the autopilot and autothrust and fly the recovery manually with expedience. This is not a situation for managed modes. It is a situation for direct, immediate, manual control.
The automation that is helpful in cruise and on the approach is not equipped to manage a developed unusual attitude. The crew that hesitates at that moment — that looks for an automated solution to a situation requiring direct manual intervention — has inverted the logic of automation selection entirely. Here, the appropriate level is no automation at all. The question still applies: will this level of automation reduce workload in these conditions? In a developed unusual attitude, the answer is unambiguously no.
This is the full range of the behaviour. From the routine cruise where full automation management is the clear choice, through the dynamic vectoring environment where selected modes may serve better, to the turbulent approach where autothrust judgement is required, to the upset scenario where everything comes off and the crew flies. The principle is identical throughout. The question is always the same. Only the answer changes.
Judgment, experience, and training determine the answer. The question itself — will this automation level reduce workload in these conditions — has to become habitual.
Briefing the Decision
Automation level selection is most often treated as a reactive decision — something chosen in the moment when the situation demands it. The behaviour at its higher levels of development treats it as something that can be partially anticipated and briefed.
Before a demanding approach in turbulence, the question of autothrust management is worth a sentence. Before a departure into a busy terminal environment with likely re-clearances, the question of when LNAV will be superseded by vectors is worth naming. Before any phase where the automation picture is likely to be dynamic, a brief agreement on the crew's automation strategy — including at what point they will step down the ladder, and who will call it — converts a reactive decision into a prepared one.
That preparation does not remove the need for in-flight judgement. It does mean that when the moment arrives, the crew is not making the decision for the first time under pressure. The appropriate level has already been considered. The trigger has already been agreed. The response is rehearsed rather than improvised.
On the Line
High Performance Pilot structures your development of Selects Appropriate Level of Automation across three levels — Foundation, Proficient, and Mastery. Each session takes minutes. The development happens on every flight. Free to start.
Start Free — highperformancepilot.com