3.1 Data classification: a questionnaire in hindsight

Self-reports on situational variables and events in hindsight contain data biases and confounding variables without structural or statistical control. Ambiguous questioning collects spurious piecemeal information on incapacitation thoughts, fatigue and physiological needs. The collected data refer to similar but different factual events that lie unsystematically distributed in the past for each participant, unevenly frequent in varying situational contexts within and between the pilots. Thus, the questionnaire co-assesses memory decay, recall and availability bias by leading questions ambiguously evoking remembrance of past events. Furthermore, these self-reports cannot measure human behaviour reliably due to a social desirability bias [Reference de Winter, Dodou and Hancock74, Reference O’Connor, Buttrey, O’Dea and Kennedy75], because the contents relate to flight safety, professional ethos, rules and consequently potential sanctions. As such, they are intrusive and elicit concerns affecting replies [Reference Yan76]. Furthermore, situational and environmental factors of daily occupational (on-duty, off-duty) characteristics, temporary health symptoms and private life stages affect the replies at the moment of participation. An objective and reliable assessment of the psychological constructs of interest during flight would use an in-field structured questionnaire administered at a fixed point in time directly after the corresponding events or timespan. A pilot of a target population (whom to count) would engage in structured journal-keeping after flight with the same questionnaire (how to count) for a distinct time span (when to count), and include at least some well-acknowledged scales (how to compare). Simultaneously, it would collect possible confounding variables: information on environmental factors such as flight duration, time of day, eating and drinking habits before and during flight and personal countermeasures taken that might influence physiological needs. This procedure would reliably describe how often and how long lavatories were used, how long durations were in-between, and which situational variables moderate or mediate the effects.

In sum, many of the items suffered from inaccuracies and ambiguity in wording, leading to various possibilities for individual interpretation [Reference Osterlind77]. There are at least 10 double-barrelled questions containing more than one statement to reply to. In addition, the frame of reference is often missing, leaving room for different interpretations of the setting. Associated situations stretch from the last flight to short-, medium- and long-haul flights, or business-related scenes like an overnight stay at a hotel, breakfast, dinner, free time and sightseeing. In the end, this procedure mismatches epidemiological assessments, the aim it implicitly pursues regarding health-related phenomena, but does not achieve.

3.2 Sampling in the thousands: confusion with epidemiology

Applying availability sampling, a type of non-probability sampling [Reference Babbie78], denotes a way unsupported by probability theory for epidemiological purposes. It can only fulfil an exploratory research purpose by skewing the European pilot population to available and willing-to-reply participants. The result is a subpopulation. Sample sizes of thousands in psychological research are only common when standardising a thoroughly developed scale in clinical psychology or applied cognition after many cycles of research. The huge sample size cannot artificially increase the scientific quality. Scientific data collections to investigate aeromedical issues usually rely on official sources like incident and accident reports of aviation authorities [Reference DeJohn, Wolbrink and Larcher79, Reference Evans and Radcliffe80], which are less biased because they are collected after flight. Sampling in the thousands inevitably requires scientific reasoning because sampling is a careful scientific decision and process, again relying on its own methods [Reference Lohr81].

The rationale and rigorous formal method of a survey in the engineering sciences has been decontextualised and misapplied through its transposition to human participants in the context of behavioural sciences. This practice resulted in a sociological opinion poll assessing individual situational attitudes instead of epidemiological or content-related data. In engineering, a survey refers to a systematic literature review answering a research question by applying a search strategy on digital libraries, clear inclusion criteria, procedures, quality assessment checklists, data extraction strategies and a synthesis [Reference Kitchenham and Charters82]. Even in psychology, a literature review refers to ‘conducting surveys of previously published material’ (Ref. [Reference Cooper, Hedges, Valentine, Cooper, Hedges and Valentine67]: p. 4), which cannot be applied to human participants on factual topics in confusion and hindsight as if they were a static database. In contrast, the method survey in behavioural sciences methodologically consists of sampling, questionnaire design and a resulting scale to measure psychological variables to objectively, reliably and validly obtain empirical results from human participants.

Extrapolating from mass data to prevalence, frequencies and characteristics of behavioural variables is the purpose of an entire scientific discipline spanning medicine and psychology called epidemiology – the science that studies disease occurrence and health states in human populations [Reference Lash, VanderWeele, Haneuse and Rothman83]. It collects data by using a complete range of its own methods to yield objective data on diseases, applies them carefully and is not as intuitive as it may seem. This professional exercise in methodologically sophisticated counting within groups without bias on whom to count, how to count them, when to count them and how to compare them comes with several complexities that all have to be addressed specifically. However, a form of intuition appears to have guided constructing the survey to thousands of pilots. Only a representative or carefully selective sampling of preferably active pilots who are healthy, free of symptoms, syndromes or diagnosed diseasesFootnote 2 could tell about averages in the pilot population.

3.3 Preparing data for analysis: cherry picking

How were these data prepared for analysis? The preparation neither applied best scientific practices [Reference Osborne86] nor followed a rationale that can be deemed objective. Subjectively tainted exclusion criteria (Table 2) instead of scientific inclusion criteria misled the process into arbitrarily selecting data, question by question, resulting in several different samples. A bottom-up approach is apparent in looking first at the data, then selecting results according to a sense of utility to reply to research questions without testable hypotheses. That is not exploratory research but rather reversing the rationale of science, also known as cherry-picking [Reference Murphy and Aguinis87], ‘searching through data involving alternative measures or samples to find the results that offer the strongest possible support for a particular […] research question’. The data preparation created further sampling issues described as follows: ‘Sampling is not mere substitution of a partial coverage for a total coverage. Sampling is the science and art of controlling and measuring the reliability of useful statistical information through the theory of probability’ (Ref. [Reference Deming88], p. 2).

Table 2.

Inclusion/exclusion criteria in data preparation for analysis

Table 2. Long description

A table with two rows and five columns detailing inclusion and exclusion criteria for data preparation in analysis. The columns are labeled ID, Topic, N before (base sample), Selection: Inclusion/Exclusion criteria, and n after (sub-sample). Row 1: ID D 7, Topic Breaks due to physiological needs, N before 6645 incomplete, Selection Exclusion of responses due to unrealistic value, nonsense, not meeting [unknown] criteria Inclusion of all complete replies to each question per single case decision Inclusion of ‘useful responses’, n after n 1 6092, n 2 6300, n 3 N A, n 4 N A. Row 2: ID D 5, Topic Pilot incapacitation management (including fatigue), N before n 1999 complete, Selection Random selection of 340 responses on sleep inertia out of an unspecified subsample, n after Quantitative n 6300 Qualitative on sleep inertia n 340.

Furthermore, four commercial pilots who had experience in types of operations that use bladder relief solutions during flight (glider and military flight operations, general aviation) were interviewed about their physiological needs during flight (D-7, pp. 18–19) and the solutions proposed by the two OEMs. The procedure, rationale and way of questioning are unreported. Unscientific reporting impedes further transparent evaluation. Interpretation is not possible, although it had been undertaken.

3.4 Deviation from established scientific and industry methods

As introduced in Section 1.2, the concepts of eMCO and SPO, lacking scientific and industry-standard risk assessment methods, [46] cannot assess risk. Examples to iteratively progress toward it in HCD were task analyses, formal process charting methods, interface analysis methods, incident and accident analyses, design methods and prototyping [Reference Diaper and Stanton89–Reference Salmon, Stanton, Walker, Hulme, Goodie, Thompson and Read92]. The SAE International guidelines [46, 47] operate in tandem with the international standards of hardware and software design established by the Radio Technical Commission for Aeronautics (RTCA) and the European Organisation for Civil Aviation Equipment (EUROCAE) [93, 94]. They are essential when flight safety is to be addressed in comparison to its current status. A workshop attendee affiliated with Dassault recommended the use of ‘methodology that is recognised by the industry such as bow-ties’ and ‘during a deeper level of analysis, some academic recognised and demonstrated tools’(Ref. [95] D-COM-3.1, p. 13). The technical project lead from NLR passed over the factual arguments in the requests to include such methods by saying that bow-ties were mathematically identical with the scenarios of the risk assessment. This is an apples-and-oranges comparison because bow-ties [96] and the flight simulations are two different methods using their own rationales, data and settings. In contradiction, the technical project lead informed the audience straight ahead about ‘the project doesn’t have the needed level of detail to do the specific risk assessment’, after the Dassault attendee had asked ‘how the project aims to provide a valuable result concerning the fact the project is doing generic analyses for concepts that are design dependent?’ This scene describes how a gap between the theoretical research foundation and empirical practice has widened, lacking the previous results on RCO from science and established methods, simultaneously having isolated RCO and the topics of medical issues from the same. The following questions arise accordingly: How were physiological needs and incapacitation identified as primary issues to end up in a questionnaire? What purpose could these data serve?

Additional workshops were project meetings with stakeholders that did not apply a scientific design specification method of human factors. Thus, the workshops guided the project into arbitrary subjective directions by leading communication. They unsystematically listed human conditions (Ref. [97] D-2.1: June 2023 N = 6) and system-failure conditions (Ref. [98] D-3: N = 2) that affect flight safety in eMCO. Selected from them, fatigue, sleep inertia, pilot incapacitation and physiological needs became subject to general literature reviews (Ref. [99] D-4, D-5, D-6, D-7). In particular, sleep inertia lacks epidemiological studies beyond aviation [Reference Sauvet, Beauchamps and Cabon100] and forbids recommending 35 minutes as a reference to EASA based on its dedicated review and the survey (D-4). In addition, the consortium proceeded with flight simulations to evaluate the effects of crewing and nominal operations, failure conditions (both in Ref. [101] D-2.6), sleep inertia and boredom (Ref. [102] D-6.2) on variables of interest for RCO under current CAT operations. By including their flight simulators without prototypes of new systems, the experimental design can only compare within its design and technological setting. In conclusion, the term denoting project achievements (Ref. [103] D-8, Ref. [104] D-9) mismatches risk because it did not provide a general safety margin as ‘reachable level of safety’ (D-2.6, p. 3) by method. The empirical flight simulations (D-2.6, D-6.2) are beyond the present scope and would justify their own analysis. A preliminary selective hazard identification would match the venture in the context of a system safety process more closely [Reference Ericson105]. The health-related data from the questionnaire on pilot states cannot contribute to any research progress on RCO because of the two issues. Even if the flight simulator settings were RCO, the health data lack scientific integrity, as shown.

4.1 Informed consent, transparency and deceptive potential

The informed consent form excluded the ConOps, explaining how pilot incapacitation, fatigue, sleep inertia and physiological needs were solved. Instead, it introduced eMCO and SPO by their terms, spelling out the acronyms and announcing it had been one survey amongst others that researchers of an EASA-funded project used to ‘assess issues and the feasibility of the implementation of both eMCO and SiPO in the EU regulatory framework’ (D-5, p. 40). Then, it listed the four health-related issues as topics. In addition, interviews, workshops and experiments gathered ‘more detailed information and ideas from the pilot communities’ with the survey on opinions. The questionnaire consisted of ‘mainly multiple-choice questions’ and some open questions, totalling 31 items. Voluntariness, anonymity of participation, confidential handling of data according to the General Data Protection Regulation (GDPR) and the option to drop out at any time had been listed. After concluding the study, all data would be disposed of, whereas data handling of dropouts (69.92%) had been omitted. A yes/no item collected agreement to and understanding of this informed consent. At the bottom, an impersonal email address utilising the project acronym without reference to the performing and liable unit was provided, in exchange for their own email address when taking up communications. The researchers operating as external contractors of EASA remained unspecified regarding affiliations. The funder is distinct from the performing organisation. Without referencing liable entities, enforcing GDPR is legally impossible because the domain ‘eMCO-SiPO’ is neither a registered organisation nor an individual. The data handling claim is legally void. However, an unregistered email domain with an incomprehensible corresponding data storage would violate core provisions of GDPR.

When reading it, pilots have two choices. Taking part gives the opportunity to obtain knowledge about why the ‘feasibility of the implementation’ of RCO calls for disclosing sensitive health-related information and events at the regulatory level. This term simultaneously implies a too-mature state of eMCO in contradiction to its current scientific progress. Implementation, ‘the act of creating a reality from a specification’ [46], consists of four elements during the development of aircraft systems [47]: information flow between the system process, which includes development assurance and safety assessment processes, to and from the item-level processes; item design/build; item integration and verification; aircraft/system integration and verification. Feasibility refers to whether something is possible, here an unspecified implementation. Neither was part of the project. These activities would be performed much later in the design process when mature prototypes are built to address their compliance with certification requirements. Thus, the text implicitly contains certification aspirations, as does the invitation [66]. It consequently misleads pilots into taking part by describing a regulatory-level developmental stage that does not exist yet. Not taking part denies pilots access to information about what is going on concerning their occupational future. Reading it confronts pilots with an ethical dilemma regarding their decision to take part.

This practice neither equals a cover story from social psychology to obtain judgements about something that too many facts would bias, nor parsimoniously introduces a topic or keeps details confidential. Social psychology makes use of deceptive cover stories to increase the interest of participants by providing a credible rationale for collecting data and demanding judgements in case of ethical justification [Reference Wilson, Westgate, Aronson, Gilbert, Fiske, Finkel, Finkel and Mendes110]. Here, missing an ethical exit by debriefing created a coercive environment when approaching pilots [Reference Mumford, Higgs, Gujar, Panicker and Stanley111]. High dropouts (69.92%) indicate that the narrative lacked something essential to capture attention, such as plausibility. Until today, the invitation and the procedure have persisted as an ethical dilemma for the participants because a professional debriefing [Reference Gluck, Swartzman and Miller112] is missing according to the reports. Pilots would have to search on their own to get to know the circumstances. SPO and eMCO were put on hold for an indefinite period and listed among finished projects [55], without explaining their scientific, ethical and industrial impact. Demanding information on events that have jeopardised flight safety in hindsight simultaneously creates an unofficial operational safety record outside the airline and an incident investigation board. These data encompass medical records and fitness to fly.

Hiding and giving unspecific spurious piecemeal information [Reference Baumrind113] on the ConOps makes the research contents and their purpose non-transparent by omission. The ethically critical and hidden suggestions of how physiological needs are imagined to be solved in eMCO take away the current freedom of the pilot over a natural and daily function of the human body: using a lavatory based on their situational and, crucially, healthy physiological needs. Put simply, to ‘pee’ or not to ‘pee’, when, and for how long is not to be negotiated because it would create another lasting coercive environment. The imagination can create suppressing thoughts and affect pilot cognition by creating worries about it. Although these are unlikely to develop into any psychological disorder, they represent the core underlying the specific-phobic component of urophobia, paruresis and parcopresis [Reference Hammelstein and Soifer114–Reference Kuoch, Meyer, Austin and Knowles116], colloquially the fear to ‘pee’ and ‘poo’. An indignation as an emotional reaction to perceived injustice [Reference Kals, Strubel, Maes, Zieris and Bruckbauer117], alongside incomprehension, could be a reason why 69.92% did not complete the survey. The suggestions on solving physiological needs behind it would scrutinise their own responsibility and consequently human autonomy over bladder, bowel and gastric functions. The four items at the bottom of the survey might have implicitly communicated it by anticipation. Furthermore, D-7 withholds the names of the two OEMs, detaching them from their suggestions about how physiological needs were solved.

The coercive environment unfolds in a top-down direction. Snowball-style sampling unsystematically distributed the survey to pilots via airlines and pilot unions. It fulfils the psychological criteria of a usual spam email on the internet: awkward prose, abnormal message structure and an implausible premise in a questionable attempt to obtain sensitive information by asserting trustworthiness [Reference Desolda, Ferro, Marrella, Catarci and Costabile118, Reference Williams, Sarno, Lewis, Shoss, Neider and Bohil119]. This spam-like email without reference to individuals collects data from the masses. However, the airlines and unions, as distributors acting on behalf of EASA, made it appear a respectable practice, although it was not.

This placed thousands of pilots in the plight of worrying about not only a future unspecified concept that might endanger flight safety, but also their personal occupational and private future and their daily physiological needs. Pilots would never have considered the needs in such detail had they not received the email. Regarding a visible cognitive preoccupation of the human participants, the unethical mass outreach lines up differently next to prominent experiments in the history of psychology that potentially affected the well-being of participants after a perceived authority had demanded ethically doubtful choices [Reference Milgram120]. Such a procedure exposed the pilots to an ethical dilemma. The driving force is the emotional and professional attachment of the pilots to their occupation and livelihood. The potential psychological harm [Reference Nebeker, Bartlett Ellis, Arigo, Panicker and Stanley121] in excessive preoccupation with worries about RCO is already evident in the worldwide emotional reaction of pilots visible in media [Reference Broderick122], public discussion and an increase in lobbying activities [Reference Reed16]. A questionnaire like this is prone to worsen negative psychological reactions that some pilots have shown already. Whether and how these worries become clinically relevant remains a separate issue, but this can result from participating in studies with ethical shortcomings [Reference Smith and Richardson123]. Since a professional debriefing has not been accomplished yet, we provide it here by rectifying disseminated misinformation [Reference Greene, de Saint Laurent, Murphy, Prike, Hegarty and Ecker124], including speculations, and by resolving the concomitant methodological and ethical shortcomings. This professional independent debriefing safeguards the aviation community by preventing further distress or psychological harm mainly at pilot and airline levels.

4.2 Data handling

Disposing of these research data necessitates a trade-off between privacy, transparency and scientific integrity in an ethical debate. The GDPR protects participants’ individual rights. An absolute assurance of safety by unlinking vulnerable populations from their data favours disposal. In contrast, integrity, transparency and replicability of science by archiving data for funding agencies and esteeming human participation and efforts contradict disposal. Ethical research requires careful planning by balancing the reasons for anonymising data against destroying them. In this case, the informed consent form mandates data disposal, after illogically introducing the contradictory argument of anonymous participation – a measure designed to retain data. The course of data handling appears illogical. Throwing away a database whose data are anonymised represents a procedural paradox. Pilots are not considered a vulnerable population, such as undocumented immigrants, victims of violence or incarcerated individuals [Reference Findley, Ghosn and Lowe125], but the protocol handles their anonymised data as if they were. We cannot tell the true reasons underlying this unusual contradictory practice. A methodologically compromised questionnaire reduced thousands of pilots to a passive database rather than employing established methods.

4.3 Anonymity diffusing responsibility and social loafing

Diffusion of responsibility [Reference Darley, Baumeister and Vohs126] systemically characterises many operations, stakeholder involvements and the project structure from top to bottom. It starts with two aircraft manufacturers, officially unnamed to date, requesting a decision from EASA associated with flight risk: the question of flight safety for eMCO and SPO is unspecified in more detail. External contractors conducted the research on their behalf but under an acronym name. Staff, company compliance, codes of conduct are the institutions’ responsibility in case of violations, the formal context of liability in a psychological setting. This differs from most of the internally and externally sponsored research projects. They operate under the project title and acronyms, including their own corporate design. In this way, EASA are distant from, rather than outside of, responsibility in case of possible consequences from violations. In turn, the commissioned researchers remained anonymous during study conduct, although their institutions retained full operational responsibilities. Only the meetings and subsequent project reports disclosed the individual identities. Hence, individual project-related activity becomes insulated from business reputation and compliance with rules of the occupational setting. Impersonal communications without individual and institutional names shape the survey invitation and the informed consent. The email address of the informed consent documentation mirrored the diffusion of responsibility by omitting both.

Regarding research quality, peer-reviewed publishing and presenting at international conferences provide valuable feedback on methodologies from independent scientific peers. The project bypassed both mechanisms, which constitute an upward social comparison to evaluate research quality. It makes scientific knowledge on a subject and its corresponding methods more accessible to revision and cohesion inside science. Thereby, it integrates work into science by assimilation to the current state. Under free choice, upward comparison dominates with a possible lateral choice reducing differences between upward and downward, whereas the latter is avoided [Reference Gerber, Wheeler and Suls127]. Free choice by social comparisons demonstrating research independence is not applicable to the eMCO-SPO project. Outsourcing the safety framework to national institutions can theoretically preserve regulatory independence. Instead of ensuring this unbiased buffer, the political mandate of EASA was transferred to the consortium by reversing the original purpose of external contracting. Leadership was divided into project and technical lead, whereas it diverged in operational conduct. Research activities were delegated across the consortium. The business procedures failed to establish a critical legal safeguard against undue industrial influence and rather precluded the opening of a scientific forum on RCO. The project’s social setting and structure prevented social comparison by creating an isolated in-group. A subjective setting emerged on individual and systemic level visible in its exclusion from scientific debate. The resulting social setting led to a diminished sense of agency at an individual level [Reference Beyer, Sidarus, Bonicalzi and Haggard128], provoking social loafing in the group. This tendency to work with less effort in a group than alone [Reference Fishbach, Gilbert, Fiske, Finkel and Mendes129, Reference Karau and Williams130] increases with in-group size [Reference Ingham, Levinger, Graves and Peckham131, Reference Karau and Williams132]. In such a group setting, members can perform below their individual capacities.

Stakeholder meetings and workshops gathered scientifically unstructured feedback from attendees selected by invitation. Internationally renowned scientists of previous and current RCO research and its methods did not attend. It is unreported who and how attendees were selected. Expert work on RCO is avoided in citation and method selection of the project. An isolated in-group setting likewise emerged in contact with stakeholders. The most important scientific criticism was passed over, as Section 3.4 demonstrates regarding the improvements an attendee from Dassault suggested. At this meeting, an anonymised polling tool explicitly identified that a ‘clear justification’ for proceeding was required (D-COM-3.1, p. 14), but was not commented upon further. Here, responsibility diffused, and the project continued despite criticism from external stakeholders and the results of the anonymised polling. These restricted lateral social comparisons proved ineffective at integrating critical feedback into the research progression. In sum, the concept of collective narcissism and its science denial [Reference Golec de Zavala133] can plausibly explain this group behaviour and dynamics.

4.4 Social-psychological dynamics in project in-group defensiveness

The group behaviour aligns with behaviours characterising a strong collective narcissism [Reference Golec de Zavala, Dyduch-Hazar and Lantos134], under which social comparisons, even at group meetings, could not function as feedback. If they had functioned, all suggestions would have been integrated to some extent into the research course of the project. Instead, meetings operating as ends in themselves serve to reinforce a group’s internal position, prioritising visibility over substantive scientific and professional content. In social identity theory, low self-esteem can prompt out-group derogation to establish positive in-group distinctiveness and enhance self-esteem [Reference Golec de Zavala135]. The phenomenon can occur among healthy and competent individuals when over-identifying with a political mandate. It is all about the process, not about the individuals.

Collective narcissism manifests as a belief in the in-group’s superiority and a biased perception of inter-group reality [Reference Golec de Zavala, Dyduch-Hazar and Lantos134]. It exhibits a tendency to protect the in-group’s positive image, including a selective construction of the past concerning the group. The eMCO-SPO consortium sought to compress almost two decades of scientific research on RCO into two official years and seven months, from 09/2022 to 04/2025 [1], with rulemaking aspiration. A ‘risky shift’ toward investigating another subject such as selective health-related issues in an isolated survey unrelated to RCO emerges more likely in such a responsibility-diffused setting. The group dynamics can be understood as a group polarisation leading to populism within science [Reference Golec de Zavala and Keenan136].

4.5 The issue of rulemaking

The ethical issue lies in the rulemaking purpose for which sensitive data were collected from an excessively large sample, while the topic remained isolated from RCO because its construction aimed to provide an epidemiological reply. This questionnaire study is considered unethical for several reasons that show non-compliance with the ALLEA Code [106], as shown in Table 3. The survey cannot support the common project goal of informing about flight safety of RCO or related research issues; consequently, it constitutes an inefficient and invalid part of authoritative rulemaking. It would build regulatory policies on compromised data. This major issue has not yet been communicated to the participating pilots and the international aviation community.

Table 3.

Summary of methodological and ethical critiques

Table 3. Long description

Table 3: A table summarizing methodological and ethical critiques. The table includes columns for specific critiques, reasons for non-compliance with the ALLEA Code, and the impact on the survey’s validity and efficiency. Each row details a different critique, such as the excessive sample size, the isolation of the topic from RCO, and the lack of communication to participants.

6.1 The value of experts and CEOs: include their expertise by action

An alternative US approach has been having the (at-this-time incoming) head of the aviation authority reply to four precise questions. Their contents encompassed the potential benefits of RCO, impact on passengers’ safety, the maturity of available software in equal decision making of two well-trained pilots and flight safety considered against the current certification requirements [Reference Bedford63]. Although the hearing took place in a politicised environment [Reference Sheehy64], it aligns closely with the current state of science. Passing such questions on, as occurred in Europe, favours not being responsible for a decision by delaying an official statement to externals and making them responsible for the contents. On the other hand, EASA employees would have a conflict of interest. In short, simply interrogating a competent CEO or expert in a Q&A session can be more evidence-based, cost-efficient and effective than conducting a €930,000, ∼2.5-year research project. Depending on context, expenditures can vary, but involving experts is always a possible countermeasure. Such peer feedback, as the Dassault-affiliated attendee gave, can lead a research project to the correct direction and ensure the very basics.

6.2 The prevention of displacing industrial R&D to the regulator

One process detrimental to quality has shaped safety-related rulemaking, namely the diffusion of responsibility. It originated with the two OEMs approaching EASA to request a safety solution for eMCO and SPO. The prematurity of some of their essentially required technologies is known regarding this application [Reference Schmid and Stanton2, Reference Driscoll41]. This dynamic relates to outsourcing of regulatory risks before they have been rigorously evaluated within industrial R&D. However, this is one purpose of R&D and business in general: managing and mitigating risks in systems engineering. Thus, it questions which competences and capabilities are available at the two OEMs to translate the current state of science in an aerospace engineering application to practical industrial business. This raises the fundamental question: What were the institutional and operational conditions under which EASA decided to allocate public funds to a rulemaking task, which all other aviation authorities are free from in the context of the current state of research progress? The displacement of industrial R&D responsibilities carries immediate economic consequences and should be actively prevented. Notably, the practice yielded an indirect economic benefit for the two OEMs externalising R&D to the regulator, as it did for the contracted external institutions receiving public funds to conduct the work, although the purpose appears conceptually unviable against the state of RCO at that date.

6.3 Upgrading capabilities: institutional quality assurance mandates and oversight

The trajectory of the project establishment shows an urgent necessity to systematically improve whether and how R&D capabilities in research design, methods, statistical analyses and applicable guidelines and standards are employed at the OEMs, regulator and external contractors. Nevertheless, these lessons reach far deeper into each organisational structure, impacting many administrative and operative units across many hierarchical levels. The institutional mandate of quality assurance has at some point fallen short of ensuring that the very basics are utilised. An attempt to do so was visible at the Dassault-affiliated attendee and some pilots as externals at meetings (D-COM-3.1, pp. 12–13). While an R&D employee of an aircraft manufacturer is expected to apply these methods, including project management competencies, the pilots positively stand out by recognising and commenting on them. It is not a firm part of their usual professional education and specialisation. Research design in empirical sciences, trading off Bayesian against frequentist statistics, aerospace design guidelines and standards, human factors and ergonomics methods, epidemiology and questionnaire design need to be integrated into existing business structures, and when unavailable in some instances, updated.

6.4 Appreciation of stakeholder reactions: differentiate science from industrial interest

At first, EASA aggregated eMCO and SPO into a continental rulemaking task, prior to integrating detailed system parameters and foundational R&D methods. Then, ICAO successfully mitigated and corrected the industrial push toward RCO through its definitive flight safety imperative in 2025 [36]. At the same time, EASA put the topic indefinitely on hold. A proposed framework proved to be a superseded draft. It could not withstand the pressure of methodological, operational, and technical realities.

Whether pilot activism has ultimately reserved the pilots a place at the table during European rulemaking is unclear, but it likely contributed to the discourse, given that the international experts on RCO were absent. Although science must not be conflated with industrial lobbying, the pilots decisively provided the latest critiques on RCO and the rulemaking framework in the meetings. RCO experts could not do so even if they had been aware of the course of the topic from industry into rulemaking. Finally, pilot union activism has been an effective contributor to project-internal safety deliberations, backed by an industry affiliate and the current state of science. Almost one decade of resourceful pilot union activities is not a defensive reaction, but an in-field safety feedback to research. As such, it constitutes an essential part of public discourse on safety-related topics that takes up research-practice gaps from an operator perspective. However, the fact that an industry affiliate would have provided the most appropriate methods to establish safety metrics in a rulemaking project turns its purpose away from the political mandate. This scenario paradoxically implies that public rulemaking is contextually dependent on external industry competence to execute its tasks reliably. For this reason, research integrity deserves an explicit position in a regulatory debate, in particular when a project and its group dynamics are documented in detail. This procedural transparency allows other academic disciplines, such as organisational psychology, to take it up as a research base to identify, explain, and optimise operational aspects that caused the issues.

Looking out and forward to future flight safety, the foundational glitch in European rulemaking is likely to persist with global reach. Changing the name of R&D relating to eMCO or SPO to imaginable terms such as ‘Autonomous Cruise Support’ or ‘Cockpit Crew Assist’ will never prevent it from being permanently associated with RCO, as smart system architects can creatively rename automation or autonomy. Rebranding a failed project through terminology or marketing cannot wipe out aviation history.