The Regulator Is Now the Bottleneck

EASA has published its first regulatory proposal for artificial intelligence in aviation, a concept document meant to set the terms for how AI-enabled systems get certified across civil aerospace AI applications in civil aerospace. For manufacturers and airlines building AI into flight systems, maintenance diagnostics, and ground operations, this is the long-awaited signal that a certification pathway exists at all. It is not, however, a signal that the pathway is ready to carry volume.

ASD Europe, the trade body representing the continent’s aerospace and defense primes, has made the capacity problem explicit. Certifying AI-enhanced systems at the scale the industry is building toward requires EASA to develop new standards, and that requires funding and specialized expertise the agency does not yet have in place Bringing EASA at scale for European aeronautics. This is a different problem than whether AI systems meet safety bars. It is whether the institution responsible for verifying that they do has the staff and budget to process the submissions.

This distinction matters for anyone sequencing a product roadmap or a procurement decision against a certification date. A manufacturer can have a technically compliant AI system, a complete DO-178C-aligned verification package, and a cooperative relationship with notified bodies, and still face a queue determined by how many qualified EASA reviewers exist to evaluate machine learning assurance cases. Regulatory readiness and regulatory throughput are not the same constraint, and only one of them is inside a vendor’s control.

The UK’s recent £600 million AI investment offers a useful contrast. A portion, £60 million, is earmarked for a Sovereign AI Research and Development Procurement Scheme spanning defence, aerospace, and engineering applications UK unveils £600M AI investment for national security. That is capital aimed at building sovereign capability on the industry side. It does nothing to expand certification capacity on the regulator side unless some of it flows toward the agencies that will eventually have to sign off on what gets built. Governments funding AI development without proportionally funding AI oversight capacity are building a queue, not a pipeline.

There is a second-order risk here that touches program governance directly. UL’s guidance on agentic AI in aerospace and defense engineering warns that autonomous action without process governance creates mission, quality, and program risk Governed AI processes for aerospace and defense. Internal governance discipline becomes more, not less, important when the external regulator is capacity-constrained. Programs that cannot rely on fast external review need their own assurance processes to be airtight, because there is no backstop timeline to absorb late-stage surprises.

For compliance and program leaders, the practical takeaway is to stop treating “EASA will have a framework” and “EASA can execute that framework at the pace we need” as the same assumption. Build certification lead time into program risk registers as its own line item, track EASA’s staffing and standards development separately from your own technical milestones, and weight vendor selection toward teams that can produce assurance documentation clean enough to move through a thin review pipeline without rework.

The rules are coming. The question every aerospace program now has to answer is how many people will be available to apply them.

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 core argument—that regulatory throughput is a distinct constraint from regulatory readiness—is coherent and worth saying, but the piece leans heavily on a Facebook post and LinkedIn update as prim
Source & Claim Verification	glm-5.3-flash-fp4 · local	held. The draft’s foundational claim — that EASA has published its first regulatory proposal for AI in aviation — is cited only to a Facebook post whose fetched content is a truncated title with no substant
Regulatory & Framework Fidelity	Mistral	cleared. The briefing accurately reflects DO-178C implications for AI in aerospace but does not substantively address ISO 42001, EU AI Act, NIST AI RMF, or other sector-specific frameworks (HIPAA, CLIA, NERC C
Technical Accuracy	Llama	cleared. The article accurately highlights the regulatory bottleneck in AI certification for aerospace, but one cited source appears unrelated to the topic.
Bias, Balance & Hype Control	Gemini	cleared. The briefing effectively identifies and addresses the core concern of regulatory capacity as a bottleneck, using relevant sources to support its claims and offering actionable advice.
Novelty & Non-Duplication	Grok	cleared. EASA’s first AI concept paper and ASD Europe’s explicit capacity/funding shortfall are new wire facts that frame regulator throughput—not readiness—as the constraint, an angle not restated in prior av
Validation	DeepSeek	cleared. The central claim that EASA’s capacity is a bottleneck is strongly supported by the ASD Europe source explicitly citing the agency’s lack of funding and expertise to certify AI at scale.

Sources cited: 7. Validation challenges: 0. Learn how these briefings are written and verified.