Does the FAA certify AI today? Current status

The short answer, from the FAA pages we cite: there is no published FAA advisory circular or certification position paper that is a general means of compliance for airborne AI. Applicants still use existing performance-based airworthiness rules, plus a project-specific issue paper when one is needed.

Status as of 14 October 2026: the FAA page is still Roadmap for Artificial Intelligence Safety Assurance, Version I. The AI/ML technical discipline page still links that Version I resource. The product issues lists index now dates Small Airplanes Quarter 3 - September 30 (Current) (Last update: September 30, 2026). The Q3 2026 Small Airplane Issues List still lists Artificial Intelligence Software as an Emerging Technology/Issue with the same coordinate-with-Policy-and-Standards wording as Q2. We searched faa.gov again for a published AI/ML Certification Position Paper or Advisory Circular and did not find one. An older FAA_AI_ML_Discipline_page.pdf path now 404s; the living discipline page is the HTML STEP URL above.

What the roadmap is, and is not

The Roadmap for Artificial Intelligence Safety Assurance, Version I says its primary purpose is a path to assuring the safety of AI in aircraft and related systems for inflight operations. It is a living document the FAA plans to update periodically. An FAA advisory-committee presentation dates the first official release to August 2024.

The same document says the treatment of the ethical use of AI is outside its scope. It is not an advisory circular and does not publish a compliance checklist. See the FAA roadmap explainer and the framework page.

How certification is described today

For an aircraft that already has a standard airworthiness certificate, the roadmap says the aircraft must be certified again to incorporate AI. The FAA has relevant regulations for systems and equipment that require the designer to show the equipment performs its intended function and that failures are mitigated to an acceptable level. Those performance-based regulations are expected to be adequate for the safety of learned AI implementations.

The same section then says existing software and complex-hardware guidance is not adequate to address learned AI implementations, because the traceability of software or complex hardware cannot be extended to an AI system developed through data. The questions the roadmap poses are what is unique to AI/ML that cannot be addressed with existing regulations and guidance, and whether existing regulations and guidance should be modified or new regulations, guidance or special conditions need to be developed.

The FAA’s AI/ML technical discipline page says aviation research focuses on measuring the functionality and performance of AI systems in accordance with the certification framework. That is a research and policy-support statement, not a published means of compliance.

The live applicant instruction we can still cite

The Q3 2026 Small Airplane Issues List, dated 30 September 2026, lists Artificial Intelligence Software (item A-1601) as an Emerging Technology/Issue. It tells applicants to coordinate with the Policy and Standards Division if the applicant is planning to develop any artificial intelligence or machine learning based technology in their design. A means of compliance issue paper may be needed. The Q2 2026 list and the Q1 2026 list use the same wording. The Q3 HTML slug is small_airplane_issue_list_2026q3 (issue, singular); Q2 uses issues_list.

That is the current, dated FAA instruction we found for a certification project: coordinate early, and expect a project-specific issue paper. It is not a type-certification standard for AI.

What the roadmap scheduled, and what we did not find

The roadmap’s milestone table lists Certification Position Paper(s) (CPPs) for 1Q2026, to document stable policy related to specific use cases of AI/ML after several project specific issue papers. Advisory Circular(s) and the SAE G-34 / EUROCAE learned-AI standard are marked TBD. A Policy Memo on AI/ML used in the development of airborne systems is listed as 4Q2024 (Draft).

We searched faa.gov on 14 October 2026 for a published AI/ML Certification Position Paper or Advisory Circular and did not find one. The September 2026 Small Airplane Issues List still points to an issue paper, not to a CPP. We do not claim those milestones were or were not met beyond what those pages show.

Learned versus learning AI

The roadmap distinguishes static (learned) from dynamic (learning) implementations. For learned AI, each new version is subjected to safety assurance. Systems that continue to learn in the operating environment are ongoing research topics. Safety assurance methods for learning AI is a lower priority.

The FAA is supporting the development of voluntary consensus standards, including possible extension of DO-178C, DO-330 and related documents and new AI standards in SAE G-34. See DO-178C and machine learning gaps and How to map an AI model to DO-178C / DO-330 tool qualification.

Questions

Does the FAA certify AI as a product category today?

No page we cite says that. The FAA certifies aircraft, engines and appliances against existing regulations. The roadmap says an aircraft must be certified again to incorporate AI, using project-specific issue papers and, where they exist, special conditions.

Is there an FAA advisory circular for airborne AI?

The roadmap lists Advisory Circular(s) as TBD. We did not find a published AC on faa.gov as of 14 October 2026.

Can an applicant wait for SAE G-34 / ED-324?

The FAA anticipates that industry standard to be appropriate for learned AI safety assurance, with a TBD date. The current applicant instruction we cite is still coordinate with Policy and Standards and expect an issue paper.