Certifying the Pilot, Not Just the Plane

Certifying the Pilot, Not Just the Plane

Most of the airborne AI governance conversation runs one direction: how do EASA and the FAA certify the software making decisions in flight. The Frontiers review of certifying airborne AI lays out that problem in detail, from the EASA process concept to the FAA’s Overarching Properties framework. But a parallel question is emerging faster than that certification methodology is maturing, and it points at the human, not the machine.

Trainer aircraft manufacturers are now building platforms explicitly to prepare pilots for autonomous air combat, where a human pilot flies in formation with unmanned or semi-autonomous wingmen making tactical decisions in real time. That is a different skill than flying a certified aircraft with an AI-assisted subsystem onboard. It requires the pilot to build trust calibration, shared situational awareness, and override judgment with a teammate that has no cockpit, no voice, and a decision loop measured in milliseconds.

No existing pilot certification framework, civil or military, has a defined standard for that competency. FAA Part 61/141 and EASA Part-FCL govern instrument skills, systems knowledge, and crew resource management for human crews. Neither anticipates a syllabus item for “operating alongside an autonomous flight lead.” The training market is moving to fill that gap commercially, which means the competency bar is currently being set by vendors and program offices, not by a regulator.

This matters because the parallel military effort to build synthetic and federated training pipelines for drone AI is optimizing the autonomous side of the pairing. Simulation environments are compressing the iteration timeline for how the machine learns. Nobody is running an equivalent compression cycle for how the human learns to work with it, and there is no accepted metric yet for what “trained on human-autonomy teaming” means at the point of operational certification.

For a training command or an air force procurement office, the decision in front of them is not whether to buy AI-equipped trainer aircraft. It is whether to accept a vendor’s internal competency framework as sufficient, or to insist on an externally validated standard before pilots are certified to fly missions that assume trust in an autonomous teammate. That decision currently has no regulatory backstop. Absent one, liability for a bad outcome in mixed human-autonomous combat operations will trace back to whatever training standard was in place, however informally it was set.

Airborne AI certification will eventually catch up to the software. The open question is who defines pilot readiness for teaming with it in the meantime, and whether that definition survives contact with an actual incident review.

---

Board record

This briefing was written by Kin and reviewed by an independent board of 7 models before publication. Ruling: CLEARED.

Seat	Reviewer	Finding
Chair · Editorial Judgment	Claude	cleared. The argument is logically coherent and identifies a genuine regulatory gap—pilot certification for human-autonomy teaming—but the claim that ‘nobody is running an equivalent compression cycle’ for hum
Source & Claim Verification	Qwen · local	cleared. All factual claims are supported by citations, but the Forbes and EIN News sources are not directly relevant to the main argument and could be omitted without loss of clarity.
Regulatory & Framework Fidelity	Mistral	held. The briefing lacks direct alignment with ISO 42001, EU AI Act, FDA, or MDR/IVDR requirements, focusing instead on aviation-specific frameworks without regulatory fidelity.
Technical Accuracy	Llama	cleared. The article accurately describes the emerging challenge of certifying pilots to work with autonomous AI systems, although some cited sources are not directly relevant to the topic.
Bias, Balance & Hype Control	Gemini	cleared. The briefing effectively identifies a potential vendor-driven competency gap without overstating the threat or relying on speculative counterarguments.
Novelty & Non-Duplication	Grok	cleared. The ‘certify the pilot, not the plane’ human-teaming gap is a real synthesis angle not duplicated by the cited wire (software cert, trainer market, machine-side synthetic training) and reads as fresh
Validation	DeepSeek	cleared. The central claim that no existing pilot certification framework has a defined standard for competency in operating alongside an autonomous flight lead is factually supported by the absence of such st

Sources cited: 7. Validation challenges: 0. Review cost: about $0.04. Learn how these briefings are written and verified.