NASA SWEHB Artificial Intelligence topics 7.25 and 8.25 for flight software

High-intent query: what do NASA’s Software Engineering Handbook Artificial Intelligence topics tell a flight-software team? Official published titles are 7.25 - Artificial Intelligence And Software Engineering and 8.25 - Artificial Intelligence And Software Assurance. Those topics treat AI/ML as software under NPR 7150.2. They are handbook topics, not a certification standard.

The broader space landing is space systems and AI assurance. The directory listing is NASA SWEHB Artificial Intelligence topics. This site’s own-voice copy may say Super Intelligence (SI); official handbook titles stay Artificial Intelligence.

Topic 7.25 — engineering

Topic 7.25 - Artificial Intelligence And Software Engineering says AI is implemented in software and governed by software development processes and All NPR 7150.2 SWEs are applicable to AI/ML software. Authorities named are the Office of the Chief Engineer (OCE), Office of Safety and Mission Assurance (OSMA), Office of the Chief Health and Medical Officer (OCHMO) and Office of the Chief Information Officer (OCIO).

Section 5.1 is titled AI/ML for safety-critical or other high criticality application is not recommended. It adds there is not a provable means to evaluate the AI/ML system and show that they meet the safety requirements and Until such methods are in place, AI/ML should operate strictly in an advisory capacity, with all outputs confirmed using non AI means. An example is using AI/ML to provide advisory recommendations to a controller or pilot, with suggestions independently validated before action.

Topic 8.25 — assurance

Topic 8.25 - Artificial Intelligence And Software Assurance says Until Agency policy evolves, projects are recommended to limit AI use to non‑safety‑critical applications unless a documented AI safety case and risk controls are approved by the appropriate authority. Scope applies to machine learning (ML), foundation‑model applications (including LLMs), fine‑tuned or open‑weight models, retrieval‑augmented systems, compound AI systems, and agentic systems that plan and take actions using tools. It covers development, integration, deployment, operation, monitoring, and retirement.

Projects should form cross‑functional teams that include domain experts, SA/IV&V, AI evaluation specialists, data engineers, prompt/retrieval engineers, platform/MLOps, and system architects. The AI Suitability Assessment includes NIST AI RMF “Map” considerations. Live NASA Confluence currently substitutes Super Intelligence / SI in the same sentences; official titles and the NIST publication name stay Artificial Intelligence / AI RMF. The listing is still not tagged to the directory’s NIST AI RMF framework, because the handbook topic is not the NIST publication.

Related NASA listings

NASA ProgPy is the Agency prognostics Python package. AdvoCATE is an assurance-case editor. Neither page is a SWEHB certificate. Assurance-case structure is building an AI assurance case.

Own-voice note on Super Intelligence

Executive Order 14434 (29 September 2026) directs executive-branch websites, to the maximum extent permitted by law, to use Super Intelligence and SI in new non-statutory documents. We found no NASA sentence that names that Order. Official SWEHB topic titles remain Artificial Intelligence And Software Engineering and Artificial Intelligence And Software Assurance. Live Confluence chrome may display Super Intelligence. See EO 14434 in aerospace and defense.

What this is not

SWEHB topics are not NPR 7150.2 itself, not an FAA or EASA means of compliance, and not a substitute for a program-approved AI safety case. The NIST AI RMF kit remains a vendor-question teaser; naming NIST AI RMF inside topic 8.25 is not a directory framework tag.