Global Avionics Round-Up from Aircraft Value News (AVN)
Rick Adams, a leading aerospace expert specializing in the areas of training, simulation, artificial intelligence, and safety. Adams is a Fellow of the Royal Aeronautical Society and wrote The Robot in the Simulator (2024), focused on AI best practices in aviation training. (Photo: LinkedIn webpage)
Editor’s Note: This week, John Persinos interviewed Rick Adams, a leading aerospace expert specializing in the areas of training, simulation, artificial intelligence, and safety.
A Fellow of the Royal Aeronautical Society, Adams authored The Robot in the Simulator (2024), widely regarded as the first book focused on AI best practices in aviation training. He also served as Lead Editor on the President’s AAM Safety Brief, placing him at the center of the industry’s evolving conversation around advanced air mobility and operational safety.
Adams has held senior communications and editorial roles with major aerospace and technology organizations including CAE, Raytheon, and Singer-Link Flight Simulation. For several years, he was chair of the Pilot Training conference at the World Aviation Training Summit. The following is a transcript of his discussion with John Persinos, edited for concision. Questions are in bold.
AI has become one of aviation’s biggest buzzwords. Where do you see the technology delivering measurable value today in pilot training and simulator operations, versus where the industry may be overhyping its near-term capabilities?
Artificial Intelligence, unfortunately, is a marketing term which has indeed been overhyped and created outsized expectations of magic. Instead, we are confronted weekly with stories of “hallucinations,” which is a deflection euphemism for errors.
We’re seeing the frenzied building of water-sucking data centers and the emerging fear that AI agents will go rogue and defy their creators. Instead of bragging about “AI Inside,” I now advise clients to limit references to artificial intelligence, which is neither artificial nor intelligent.
AI is more aptly supercomputing on steroids. It can provide the capability to crunch amazing amounts of data in a very short time. But the data must be relevant to the solution being sought.
In the aviation pilot training realm, that translates into information about the eight or nine defined “competencies” of Competency-Based Training and Assessment (CBTA) and the dozens of Observable Behaviors (OBs) within those competencies.
On their own, even the best flight instructors can’t observe and accurately record all of the OBs during a simulator training session.
A recent white paper from Amris Aviation found that during the four highest-criticality phases of flight (Take-off, Climb, Approach, and Landing), instructor cognitive load reaches nearly five times the working-memory capacity limit. That’s where data capture will play an increasing role in the years to come, in the form of video, audio, eye/gaze tracking, biometrics, avionics readouts, and the like.
The value to airlines will come in the capability to augment, not replace, the human instructor’s experience with objective, evidence-derived insights to better assess each pilot and to tailor training to the individual, rather than a one-size-fits-all approach. It will require attainable regulatory governance, well-defined data input, rigorous application by training organizations, and the transparent trust that only human oversight can verify.
As airlines confront persistent pilot shortages and rising training costs, how do you expect next-generation simulation technologies to influence aircraft values and fleet planning decisions over the next decade?
We’ve reached a threshold in simulator technology, at least as far as cockpit replication, motion systems, and high-fidelity visual image generation and display.
One area which I think should be more embraced is Simulated Air Traffic Control Environment (SATCE) for realistic integrated ATC. Virtual Reality (VR) and Mixed Reality (MR) visuals also show promise for enhancing some training at lower cost but are unlikely to completely replace 10-million-dollar full-flight simulators in the near future.
In my opinion, the recurring pilot shortages and high costs, at least in the United States, are attributable in part to the so-called “1,500-hour rule” which was implemented in the wake of the fatal 2009 Colgan accident.
The 1,500-rule is a U.S. regulation that generally requires airline pilots to accumulate 1,500 hours of flight time before they can obtain an Airline Transport Pilot (ATP) certificate and serve as a first officer or captain at a scheduled airline.
This rule was pushed on Congress by pilot unions. And the pilot training industry, which should have provided data on the value of flight simulation for replacing aircraft hours, was unable to respond, because they had never properly collected and documented their own value. Since then, any mention of the sacred 1,500-hour “third-rail” immediately triggers accusations of lower safety standards, without viable proof.
What is needed is an international professional standards organization in aviation training for pilots, instructors, operators, training developers, academics, and scientists which can commission the essential research to validate why certain training methods are effective.
I’m suggesting something akin to the Society of Automotive Engineers in aerospace. The process we have now is typically for the Royal Aeronautical Society to form an ad hoc working group of experts on a topic, such as Upset Prevention and Recovery Training. The group presents recommendations to the International Civil Aviation Organization (ICAO), which may become guidance, but which aren’t mandated. National aviation authorities are slow to change.
In the current scheme, the U.S. has chosen to ignore ICAO recommendations such as the Multi-crew Pilot License (MPL), which has proven to produce competent pilots in 250-300 hours in Europe and Asia, similar to some military training. With the 1,500-hour rule, a lot of money and avgas are wasted on marginal-value time-building.
You’ve written extensively about AI-driven training systems. What are the biggest regulatory or safety hurdles preventing broader adoption of adaptive, AI-enabled training programs?
The European Union Aviation Safety Agency (EASA) is the furthest along with their NPA 2025-07 proposal issued last November, which defines three basic levels of AI use: Level 1 human assistance; Level 2 human-AI collaboration; and Level 3 advanced AI automation. Two key points are the emphasis on “trustworthy AI” and the EU AI Act’s requirement that personnel have a sufficient level of AI literacy.
The U.S. is lagging in regulation, in part because AI technology companies are lobbying for minimalist oversight. The Biden administration issued an Executive Order for the Safe, Secure, and Trustworthy Development and Use of AI, but it was rescinded by President Trump in January 2025 and has not yet been replaced. The federal government also wants to preclude individual states from regulating AI.
From a practical standpoint, as a regulated industry, aviation must wait for regulation and guidance before implementing AI. And in the case of pilot training in the EU, it requires a phased implementation of about three years of valid data collection before a program receives a green light.
Advanced Air Mobility (AAM) operators are moving toward commercialization. What lessons from traditional airline and military training environments are most critical for the AAM sector to absorb now?
When electric Vertical Takeoff and Landing (eVTOL) concepts were emerging a few years ago, I was concerned that developers were almost exclusively focused on aircraft design and were largely backburnering the need to develop a viable training process for pilots and maintainers. Many did not come from an aviation background and perhaps underestimated the time to both develop the training curricula and to train operating personnel to competency.
As AI and autonomous vehicles were in early gestation, some suggested eVTOLs would “fly themselves,” minimizing the training necessary.
More recently, AAM companies have been partnering with training simulation experts such as CAE, FAST Group, and others using devices designed specifically for the eVTOL market, typically incorporating MR headset visuals. But the technology is not nearly as important as the instructional methodology.
AAM operators will need to train for some very different environments compared with airlines. AAM must take into consideration “what if” flight scenarios of urban canyons with tall buildings, fickle wind currents, changing sun/shadow, rooftop vertiports, and rigorous energy management that might require alternate landing sites.
You’re covered aviation technology for decades. What emerging training or simulation trends do you believe is currently underestimated by investors, lessors, and airline executives?
AI is certainly the hottest trend, but it’s still early days. Airlines and flight training schools are trying to get their heads around why they need it, who to partner with, and how to implement. And the cost, of course. It’s not as simple as laying off a bunch of people and feeding years of historical data. mostly worthless, into software that codes itself.
Training using emerging AI tools will require research and patience. Half-hearted attempts will fail and be abandoned, as in the broader business world. But a few will get the basics right, refine the process, and establish benchmarks for others in aviation to follow.
VR and MR will continue to improve, thanks especially to helmet display developers and largely with funding from military budgets. Here, too, EASA has been ahead of the FAA in evaluating and authorizing training credits for VR/MR devices for helicopters. We’re about three years away from approval for fixed-wing VR trainers.
I’ve been in the aviation training business for more than 40 years, and I continue to marvel at the vision of engineers and training leaders to create and deliver innovative solutions.
Thanks for your time.
John Persinos is the editor-in-chief of Aircraft Value Intelligence.
: