Modern aircraft have never been safer, more reliable or more automated. That success, however, creates a peculiar training problem: pilots intervene manually less often, while still being expected to perform exceptionally when automation no longer provides the expected answer. As ICAO, EASA and the wider aviation industry move from task-based instruction toward competency-based training, a fundamental question is emerging: are we training pilots to complete exercises — or to manage the unexpected?
Automation is one of commercial aviation’s greatest safety achievements.
Modern autoflight systems reduce workload, improve flight-path accuracy and allow crews to devote more capacity to monitoring, communication and decision-making. Complex airspace concepts, precise navigation and increasingly demanding operational environments would be difficult to manage without them.
The problem is therefore not automation itself.
The problem begins when a system that works extraordinarily well almost all the time creates fewer opportunities for pilots to exercise the skills required when it does not.
That distinction is important. The debate about pilot competence should not become a nostalgic argument about whether “real pilots” hand-fly aircraft. Modern airline flying requires both sophisticated automation management and strong manual flight-path management. International training standards increasingly treat them as two distinct competencies.
ICAO’s competency framework explicitly separates Aircraft Flight Path Management – Automation from Aircraft Flight Path Management – Manual Control. EASA uses the same distinction in its Evidence-Based Training framework.
The question is therefore not automation versus manual flying.
It is whether pilots remain proficient in both.
The paradox of a highly reliable cockpit
The more reliable aircraft and automation become, the less frequently crews encounter situations requiring significant manual intervention.
From a safety perspective, that is unquestionably good.
From a training perspective, however, it presents a challenge familiar to every high-reliability profession: skills that are rarely exercised can become more difficult to retrieve quickly under pressure.
IATA examined this issue through a large study of airline pilots on aircraft handling and manual flying. Among pilots who preferred keeping automation engaged, 96.1% said workload reduction was a reason. Other responses referred to improved situational awareness, fatigue and company policy.
That is an important finding because it undermines the simplistic argument that pilots should merely disconnect the autopilot more often.
Automation creates capacity.
There are circumstances in which removing it would increase workload at exactly the wrong moment.
One pilot quoted in the IATA study put the dilemma particularly well: if the aircraft is flying normally while situational awareness has been lost, immediately disconnecting automation may leave two pilots manually flying an aircraft they still do not understand.
In other situations, automation itself may be the source of confusion and reverting to a simpler mode — potentially including manual flight — may be precisely the correct response.
This is not a contradiction.
It is a competence.
EASA is already training for automation surprise
European regulation gives a useful indication of how seriously the issue is taken.
Current EASA guidance requires automation training to address “mode awareness, automation surprises and over-reliance”, including false confidence and complacency. When training takes place in a flight simulation training device, EASA says it should include automation surprises originating both from the system and from pilot actions.
That language matters.
A pilot does not merely need to know how the autopilot or flight-management system works. He or she must maintain a mental model of what the system is doing, recognise when aircraft behaviour diverges from expectation, and choose an appropriate level of automation.
Sometimes that answer will be more automation.
Sometimes less.
Sometimes none.
This is why the phrase automation management is more useful than automation dependency. A competent crew must be able to move up and down the automation hierarchy deliberately rather than treating maximum automation or manual flight as inherently superior.
Airbus has long expressed a similar philosophy in its operational guidance: use the appropriate level of automation, understand what the aircraft is doing and take action when it does not behave as expected.
The objective is not to defeat automation.
It is to remain ahead of it.
Recognition may matter more than the manoeuvre
This leads to a deeper problem with traditional training.
Training a pilot to execute an engine failure, rejected take-off or unreliable-airspeed procedure is necessary. But knowing how to complete a known exercise is not necessarily the same skill as recognising an ambiguous problem in an aircraft that was functioning normally moments earlier.
The distinction is fundamental:
Knowing how to recover from a failure is not the same as recognising an unexpected failure under pressure.
Traditional recurrent training has often been structured around defined events and manoeuvres. The crew enters a simulator knowing that abnormalities will occur. Depending on the programme, pilots may even have a reasonable idea of which exercises are likely.
Operational reality provides no such courtesy.
The first challenge may be uncertainty.
What failed?
Is the automation reacting correctly to another failure?
Are the indications reliable?
Should the crew intervene immediately or preserve the current aircraft state while building situational awareness?
And how does human performance change when the event is genuinely unexpected?
Flight Safety Foundation has repeatedly examined startle and surprise as operational issues. In September 2026, its president, Rithi Baruah, described simulator observations in which technically capable crews handled known failures precisely but struggled when “the automation did something unexpected” or when the nature of the problem became unclear.
That is precisely the environment competency-based training is intended to address.
From checking tasks to building pilots
The shift is already visible at the highest level of international aviation training policy.
ICAO published the second edition of its Manual of Evidence-Based Training, Doc 9995-1, in 2026. Meanwhile, IATA describes Competency-Based Training and Assessment, or CBTA, as an increasingly important alternative to traditional task- or hour-based approaches throughout a pilot’s career.
The difference is more important than the terminology suggests.
A conventional training question might be:
Did the pilot complete the manoeuvre correctly?
A competency-based question is broader:
What does the pilot’s performance tell us about his or her ability to manage future situations?
The established competency framework includes communication, leadership and teamwork, workload management, situation awareness, problem solving and decision making, application of procedures and knowledge — alongside both manual and automated flight-path management.
The manoeuvre has not disappeared.
It has become evidence.
An engine failure, for example, can reveal far more than whether a pilot remembers a procedure. It can expose workload management, communication, flight-path control, decision-making, prioritisation and the ability to recognise deteriorating situational awareness.
That is a fundamentally different training philosophy.
Europe has just changed the simulator equation
This debate became particularly relevant in Europe in 2026.
On 15 July, EASA introduced a new regulatory framework for Flight Simulation Training Devices, describing it as the most significant evolution of European flight-simulation qualification in decades.
The change may prove more consequential than the technical language initially suggests.
Europe is moving away from relying solely on traditional fixed simulator categories toward a capability-based qualification system built around an FSTD Capability Signature.
More importantly, EASA has introduced an optional task-to-tool methodology.
The idea is deceptively simple:
Do not start by asking which simulator category a course traditionally uses.
Start by asking what the pilot needs to learn.
Then identify which training device has the capabilities and fidelity required to achieve that training objective.
That opens the door to a very different training architecture.
Not every learning objective requires an expensive full-flight simulator.
Cockpit familiarisation, procedural flows, some decision-making exercises, systems interaction and elements of multi-crew cooperation can increasingly be trained using sophisticated lower-level devices, digital tools and extended reality.
High-fidelity simulator capacity can then be concentrated where fidelity actually matters.
The objective should not be cheaper simulation for its own sake.
It should be better allocation of training resources.
Perhaps the industry has been asking the wrong simulator question
For decades, simulator quality has often been intuitively associated with realism.
More motion.
More visual fidelity.
More accurate replication of the aircraft.
Those qualities remain indispensable for certain training objectives.
But EASA’s new philosophy effectively asks a more sophisticated question:
What capability does this particular exercise require?
Its ongoing TRAIN research project takes the issue further. EASA is investigating whether some objectives currently trained in an aircraft could instead be achieved in simulation with an equivalent or better training output. The project also examines VR, XR and mixed reality, resilience and confidence, and whether existing instructor competencies remain sufficient as training media evolve.
The project runs until January 2028.
That research could eventually challenge one of aviation training’s oldest assumptions: that training quality is closely correlated with aircraft hours.
For some competencies, actual flying will remain irreplaceable.
For others, it may not be the most effective classroom.
Airbus and Boeing are already moving in this direction
Aircraft manufacturers themselves are increasingly designing training around competencies rather than simply sequences of manoeuvres.
Airbus said in June 2026 that aviation training is moving from traditional task-based instruction toward a more adaptable, data-driven and competency-based model, combining CBTA, technologies such as VR and greater emphasis on instructor development.
The manufacturer has also previously identified manual flight-path management as one of the more difficult competencies to maintain during line operations because commercial pilots may manually fly relatively infrequently. Its training philosophy has therefore sought to move some systems and procedural learning away from valuable full-flight-simulator time, leaving greater capacity for dynamic manoeuvring and competency development.
Boeing’s current CBTA framework similarly separates Flight Path Management: Manual and Flight Path Management: Automation, alongside the human-factor competencies of decision-making, situation awareness, workload management, communication and leadership.
Its current type-rating training material explicitly includes manual handling, automation management and non-normal scenarios designed around startle and surprise.
There is therefore growing convergence among regulators, airlines and manufacturers.
The future of pilot training is not simply about adding simulator hours.
It is about using training time more intelligently.
But technology cannot solve the instructor problem
There is a danger, however, in assuming that better simulation technology automatically produces better pilots.
It does not.
A simulator can reproduce a failure.
An instructor has to understand the performance.
Why did the pilot make the decision?
Was the underlying problem technical knowledge?
A weak mental model?
Task saturation?
Confirmation bias?
Startle?
Poor communication?
A reluctance to challenge the captain?
Or was the scenario itself poorly designed?
This becomes even more important under competency-based training because instructors are no longer simply judging whether a manoeuvre falls within numerical tolerances.
They are observing behaviour and interpreting evidence.
IATA and IFALPA therefore describe instructors and evaluators as making an “essential contribution” to flight safety and having a predominant role in training effectiveness. Their joint guidance focuses specifically on identifying, training and assessing instructor competencies.
IATA’s 2026 CBTA framework now extends from pilot aptitude testing and initial licensing through command training, UPRT, instructor/evaluator development and airline operations. It also explicitly addresses instructor standardisation and inter-rater reliability — essentially the question of whether different instructors observing the same performance reach sufficiently consistent conclusions.
That may become one of the hardest parts of the transition.
Artificial intelligence may help analyse data.
XR may create scenarios cheaply.
Simulation may become more accessible.
But none of those technologies eliminates the requirement for an experienced instructor capable of understanding why a crew performed as it did.
CBTA is not automatically better training
There is another important caution.
Replacing traditional training with something labelled “competency-based” does not guarantee improvement.
IFALPA supports CBTA as a reasonable alternative but attaches an important condition: it must be properly developed, implemented and maintained.
That warning deserves attention.
A badly designed EBT or CBTA programme can simply become traditional checking with different terminology.
The real transformation requires data, appropriate scenarios, trained instructors, meaningful behavioural observations, quality control and the willingness to adapt a programme as operational evidence changes.
In other words, the value does not reside in the acronym.
It resides in the training system.
Manual flying is not the opposite of modern airmanship
The debate around automation also frequently becomes distorted by a false choice.
One side warns that pilots are losing manual skills.
The other points out — correctly — that modern transport aircraft are designed to be flown using automation and that unnecessary manual flying may increase workload.
Both can be true.
Flight Safety Foundation has warned against reducing the issue to the claim that pilots simply “don’t know how to fly when the automation disconnects.” System design, operator policy, procedures, training and pilot competence interact.
The IATA survey points in the same direction.
Pilots use automation because it is useful.
That does not remove the requirement to remain capable without it.
The professional skill lies in understanding when automation is helping, when it is no longer helping and what level of intervention the situation demands.
That is modern airmanship.
Training for the moment nobody briefed
Commercial aviation has become extraordinarily good at preventing known hazards.
Aircraft are more reliable.
Systems provide more protection.
Procedures are highly standardised.
Training devices can reproduce situations that would be unacceptable to practise in an aircraft.
The next challenge is therefore increasingly about resilience when the event does not fit neatly into the expected script.
A pilot may never encounter a serious automation anomaly during an entire career.
That would be a triumph of aviation safety.
But the industry must still prepare that pilot to recognise and manage one correctly on the day it happens.
This is the central paradox created by highly reliable automation:
the safer and more dependable the aircraft becomes, the fewer natural opportunities pilots have to practise some of the competencies needed when normality ends.
The answer is not to use automation less.
Nor is it simply to demand more manual flying.
It is to design training around the capabilities that pilots actually need: understanding automation, controlling the aircraft manually, recognising ambiguity, managing workload, communicating effectively and making sound decisions when information is incomplete.
Europe’s new simulator rules, ICAO’s evolving evidence-based training framework and the industry-wide expansion of competency-based training all point in the same direction.
The unit of pilot training is gradually changing.
It is becoming less about the number of hours completed or exercises passed.
And more about the competence demonstrated.
That may ultimately be the most important change of all.
What’s next?
The next phase will be less about proving that new training technology exists and more about proving that it produces better pilots.
Three developments are particularly worth watching:
Europe’s task-to-tool transition. EASA’s new FSTD framework will test whether training organisations can match devices more intelligently to specific learning objectives rather than defaulting to traditional simulator categories.
The TRAIN research programme. Its findings could influence how much initial pilot training needs to take place in an aircraft and where simulation, including XR, can achieve equivalent or better results.
Instructor standardisation. As assessment becomes increasingly competency-based, the quality and consistency of instructors and evaluators may become as important as the fidelity of the training devices themselves.
The smartest simulator in the world can create an unexpected event.
The real measure of training is whether the pilot is ready when it does.
This article was edited with the assistance of artificial intelligence and reviewed by the Aerospace Central Europe editorial team.
Sources: ICAO Manual of Evidence-Based Training, Doc 9995-1 (2nd Edition, 2026); EASA Regulation (EU) 2026/781, CS-FSTD Issue 1 and associated Decisions; EASA Easy Access Rules for Air Operations; EASA TRAIN research programme; IATA Aircraft Handling and Manual Flying Skills Report and 2026 CBTA Library; IATA/IFALPA Guidance Material for Instructor and Evaluator Training; IFALPA CBTA Position Paper; Flight Safety Foundation; Airbus Flight Safety and Training publications; Boeing Global Services CBTA and flight-training material.
About the Author
