CRM Training Faces Automation Shift
EASA pushes evidence-based training and ICAO flags automation dependency as crew resource management is reshaped by AI cockpit integration.
Organizational byline
Updated August 12, 2026
8 min read

Crew resource management training has been a regulatory cornerstone of commercial aviation safety for nearly five decades. What began as "cockpit resource management" in the early 1980s has evolved through five generations of methodology, absorbing lessons from accidents, advances in human factors research, and shifts in how pilots interact with automated systems. Now, a combination of regulatory reform and operational reality is pressing CRM into its most significant evolution yet — one that will affect every pilot working toward or holding an ATPL under Part 121 or equivalent operations.
The regulatory landscape in 2026
On the European side, EASA continues to implement Evidence-Based Training (EBT) across commercial air transport operators. The framework, built on Commission Implementing Regulation (EU) 2020/2036 and the associated ED Decision 2021/002/R, moves recurrent training away from a prescriptive check-list approach toward a competency-based model. CRM is not a bolt-on module in this framework — it is embedded in the assessment and training cycle. EASA's AMC1 ORO.FC.115 explicitly allows operators to substitute a compliance-based CRM approach with a competency-based one "such as evidence-based training," with the training "characterised by a performance orientation, with emphasis on standards of performance and their measurement."
In the United States, the FAA's CRM requirements apply under 14 CFR Part 121, Subparts N and O, and Part 135 under §135.330. The agency's Advisory Circular 120-51 series defines the curriculum: communication, decisionmaking, teamwork, task allocation, and situational awareness within standard operating procedures. Part 135 certificate holders must have an approved CRM training programme that includes both initial and recurrent training, and no certificate holder may use a flight crewmember or flight attendant who has not completed approved CRM initial training.
The gap between the two frameworks is narrowing. EASA's push toward competency-based assessment is influencing ICAO's broader agenda, and the FAA's own evidence-based training adoption, while slower, is moving in the same direction.
The automation dependency problem
The central operational pressure driving CRM reform is automation dependency. ICAO's Personnel Training and Licensing Panel, through its Automation Working Group, conducted a comprehensive study that reviewed 386 reports — 77 accidents and 309 major incidents — for indicators of automation dependency. The findings were stark: 36 per cent of accident cases reviewed included indicators of automation dependency, and that proportion rose to 49 per cent for accidents occurring between 2010 and 2021.
The study identified a consistent pattern across multiple high-profile accidents: pilots failed to detect autopilot or autothrottle mode changes, experienced mode confusion, delayed manual intervention when automation behaved unexpectedly, and exhibited monitoring lapses that allowed aircraft energy states to deteriorate undetected. The Asiana Airlines Flight 214 accident at San Francisco in 2013 remains a frequently cited case — the monitoring pilot failed to detect decreasing airspeed during a visual approach in a Boeing 777, a deficiency the NTSB attributed in part to insufficient CRM, cross-checking failures, and limited experience managing automation during visual approaches.
EASA's own CRM Training Implementation document identifies the issue directly: "Under normal conditions, crew rely on highly automated systems. Complacency and overreliance on automation are well known threats and monitoring and communication have become core competences for safety."
What the next generation of CRM looks like
The current evolution of CRM is moving in several directions simultaneously:
Threat and Error Management (TEM) has become the organising framework for contemporary CRM, replacing earlier models that focused primarily on communication styles. TEM teaches crews to identify external threats (weather, terrain, traffic, system failures) and internal errors (mode confusion, data entry mistakes, monitoring lapses) as distinct categories requiring different management strategies. The LAPD's airborne law enforcement programme, profiled in AirMed & Rescue, illustrates how TEM has been integrated into operational training: quarterly check rides assess TEM competencies, and ground school dedicates specific modules to distinguishing external threats from internal errors.
Resilience development is now an explicit CRM training element under EASA's AMC1 ORO.FC.115. The regulation requires that CRM training "address the main aspects of resilience development," including coping with unexpected, unusual, and stressful situations. This represents a shift from the earlier CRM focus on error avoidance toward error management — the recognition that errors cannot be entirely eliminated and that detection and recovery are equally important skills.
Automation-specific CRM is emerging as a distinct training domain. EASA's guidance now requires that CRM training include "training in the use and knowledge of automation, and in the recognition of systems and human limitations associated with the use of automation." The pilot monitoring role has been given particular attention: EASA requires training on "when it would be appropriate to intervene, if felt necessary, and how this should be done in a timely manner," with reference to structured intervention procedures in the operations manual.
Human-AI teaming is the frontier. A 2026 systematic review by Yiu et al., highlighted in a recent narrative review published in the journal Aerospace, identifies how AI is reshaping human performance monitoring in aviation. Machine learning models can now identify physiological signals of rising cognitive workload or declining situational awareness before performance visibly degrades. CAE's CORe project uses simulator data to inform training scenarios. The CRMSON app delivers CRM-based micro-interventions drawing on academically validated sources. These tools are not theoretical — they are entering the training environment now.
The institutional gap
Despite these advances, a persistent gap remains between research and operational practice. A 2026 narrative review published in MDPI Aerospace found that "many aviation operators and training organisations still lack" the institutional mechanisms needed to systematically translate emerging scientific evidence into training design and delivery. The review identified that "advances in research on teaching and learning methods and human performance are not consistently brought forward into everyday training practices."
This is not a regulatory failure — the rules exist. It is an implementation gap. EASA's three-year CRM recurrent cycle, during which operators must cover all CRM training elements and revise the programme based on management system data including CRM assessment results, requires operators to actively analyse their own safety data and feed it back into training design. Many operators treat this as a compliance exercise rather than a genuine feedback loop.
For pilots pursuing a type rating or completing a conversion course, the practical implication is that CRM is no longer a separate classroom session bolted onto technical training. Under both EASA's EB framework and the FAA's increasingly integrated approach, CRM competencies are assessed in the simulator alongside technical skills. A pilot who flies the manoeuvre correctly but fails to communicate intentions, manage workload, or monitor the pilot flying will not pass the assessment.
What to watch
- EASA rulemaking on ORO.FC: EASA has an open rulemaking task to update the AMC and GM to Subpart FC of Part-ORO for evidence-based training, which will further refine CRM requirements. The outcome will influence how operators structure their recurrent programmes over the next compliance cycle.
- FAA alignment: The FAA's own evidence-based training adoption continues, and the agency's response to ICAO's Personnel Training and Licensing Panel recommendations on automation dependency will shape future Part 121 training requirements.
- AI integration in training tools: Simulator manufacturers and training providers are incorporating AI-driven performance analytics. The question is whether regulators will recognise these tools for compliance credit, or whether they remain supplementary.
- Reduced-crew operations: Although pressure for reduced-crew operations has eased, the ICAO working group notes it is likely to remain on the long-term agenda. Any movement toward single-pilot operations would fundamentally reshape CRM, shifting it from human-human coordination to human-AI teaming.
The pilots most affected by these changes are those entering the industry now. For context on the career pipeline and training demand, our analysis of the Boeing pilot and technician outlook provides the demand-side picture, while our guide on how to become an airline pilot covers the training pathway. The mechanic and technician side of the workforce equation, which directly affects simulator availability and training throughput, is examined in our coverage of the mechanic shortage.
CRM has always evolved in response to what the cockpit demanded. The next iteration — integrating automation management, resilience, and eventually AI teaming — will require operators to treat CRM as a living programme rather than a recurrent box-ticking exercise. The regulators have built the framework. The question is whether the industry will use it.
Sources
- EASA: Evidence-Based Training (EBT)
- EASA: CRM Training Implementation
- ICAO: Pilot training improvements to address automation dependency (A40-WP/296)
- FAA AC 120-51B: Crew Resource Management Training
- 14 CFR 135.330: Crew resource management training
- AirMed & Rescue: The evolution of crew resource management training
- MDPI Aerospace: Whither CRM?—30 Years on



