We are seeking an MIT graduate research assistant (RA) to start in spring 2027 on a research project with the FAA on pilot mental health, a challenge that continues to contribute to tragic accidents. The project began in fall 2026. The project models how regulatory policy, incentive structures, and industry behaviors interact with organizational and safety culture, reporting and disclosure, and the data systems the FAA relies on. We will recommend changes to policy and practice that reduce the barriers to pilots seeking care without weakening safety oversight. The RA will take part in every stage of the analysis and learn the systems approach we use for it. We don't expect applicants to know this approach already. We will train you and work through the analysis with you.
A motivating example
One motivating example is the Germanwings crash. A co-pilot with a documented history of depression, including suicidal ideation, flew 149 people into the French Alps. His medical certificate carried a waiver that made it invalid if the depression came back. It came back several months before the crash. Reporting it was left to the pilot, who had every reason not to, and the doctors treating him had no clear rule telling them when public safety outweighed confidentiality. His employer never found out. The information existed, but nothing carried it to the people who could act on it, for reasons that were themselves well-intentioned. Many of the conditions that contributed are still in place across the industry.
A systems problem
Pilots sit inside an incentive structure that punishes honesty: disclose a mental health concern and you risk your medical certificate, and possibly your career. Meanwhile, each level of the system above the pilot needs feedback before it can act. If nothing is reported, there is nothing to trigger a response. If a response ever carries consequences, it becomes a stronger reason not to report the next time. Each round erodes the feedback, creating a vicious cycle. In an earlier project with similar goals, we found that the system's own control actions, used to direct the people inside it, were contributing to the mental health problems in other ways by wearing down people's quality of life until it became unmanageable.
What the research will do
The research will model the control structure at each level (the pilot, the airline, the regulator, and others): who can act, what they can act on, and what they can see. We will work out what feedback, control actions, and responsibilities each level needs, and find where the structure breaks down. Some factors run through all the levels. There are reasons, for example, why each part of the system struggles to see a mental health issue before it turns dangerous. For some, the structure presents a choice between protecting a career and protecting the flight. One of the research questions is how to change the structure to improve both flight safety and mental health. In this project, you will learn to model systemic factors like these, which are common to any complex system, and to develop effective solutions.
We will develop a set of constraints that have to hold at each level if mental health issues are going to be managed effectively. If you've worked with STPA before, this is familiar territory. If you haven't, we'll treat the regulator, the airlines, the pilots, and other groups as a control problem and analyze it the way you'd analyze any complex engineered system. The project suits a systems thinker who can work across the engineering and the human sides of a problem.
Timing and how to apply
The project began in fall 2026 and runs for up to three years. We are looking for an RA to join at the start of the spring 2027 semester. If you haven't taken it already, we recommend MIT's System Safety class (16.863) in the fall, which teaches the methods and concepts we'll use on this project.
To apply, email John Thomas at jthomas@mit.edu.
New to STPA? The free tutorials are a good place to start.