Strong and well-balanced article. I particularly agree with the focus on calibrated trust and the idea that pilots will evolve into system-level supervisors rather than being replaced.
One point worth expanding is that trust in AI does not begin in the flight deck it begins upstream, with how operational information is accessed, interpreted, and shared.
From an ATC perspective, many safety and efficiency decisions are constrained less by flight control automation and more by fragmented aeronautical information (AIPs, NOTAMs, procedures, weather) and the cognitive load required to reconcile it in real time.
In my view, near term AI value lies not in autonomous decision making, but in reducing cognitive friction and strengthening shared situational awareness across pilots and ATC.
It would be a practical and acceptable way to gradually introduce trustworthy AI into human operations, while reducing resistance to change.
I am currently exploring this space through a project called SARA.