Responsible By Design (RBD)
Rating for Military AI
Independent ratings for responsible innovation
What is the RBD Rating?
The RBD Rating is an independent star-rating system that provides a transparent and objective assessment of whether national military AI frameworks, industry developments, and individual military AI systems are responsible i.e. lawful, ethical, safe, and trustworthy.
Drawing on international law, recognised ethical AI principles, engineering assurance standards and governance best practice, the RBD Rating gives states, investors and industry a clear benchmark for what “responsible” military AI means in practice.
Building Responsibility into Military AI
The RBD Rating creates a common standard that brings together states, industry, and investors around the common goal of ensuring that AI in the military context is developed and used responsibly.
By embedding law, ethics, safety, and trustworthiness at every stage, the RBD Rating helps to ensure that military AI is not only technologically advanced but also legally sound, ethically defensible, and societally acceptable.
RBD Categories, Weighting & Evidence
-
Legal Compliance — Gating + 25% weight
Demonstrate conformity with IHL, national law and, where applicable, legal review requirements.
Show “lawful by design” features: distinction, proportionality, precautions, human override.
Evidence Pack: Article 36 legal review file, CONOPS/HMI controls, test reports.
-
Ethical Compliance — 25% weight
Alignment with national/ organisational AI ethics policies (DoW, NATO, MoD).
Criteria includes: human oversight, bias mitigation, explainability, accountability.
Evidence: ethics matrix, dataset documentation, ethical red-teaming results.
-
Safety (Engineering Assurance) — 25% weight
Demonstrate hazard analysis, robust engineering, and resilience under battlefield conditions.
Evidence: safety case, hazard logs, test reports, cyber-hardening validation.
-
Trustworthiness (Governance & Lifecycle) — 25% weight
Evidence of governance, lifecycle risk management, and operational readiness.
Evidence: policies, SOPs, training, incident reporting mechanisms.
How the RBD Rating system works
The RBD Rating system evaluates military AI across four categories and awards a transparent star rating from 0 to 5.
Failure to meet international or national legal requirements results in an automatic zero rating, while higher ratings reflect stronger alignment with international and national military AI standards as well as responsible design practices.
The rating system applies at three levels, nations, companies and specific military AI capabilities to show how robust their responsible AI frameworks are. By doing so, it helps states, industry, and investors build trust, demonstrate accountability, and foster innovation that is both technologically advanced and responsibly governed.
5★ – Exemplar (Gold)
4★ – Strong (Silver)
3★ – Meets Standard
2★ – Conditional (improvements required)
1★ – Below Standard
0★ – Reject (unlawful)
RBD Ratings
-
National Ratings
An independent objective assessment of national military AI frameworks RBD Rating (coming soon)
-
Industry Ratings
An independent objective assessment of individual company RBD ratings (coming soon)
-
Capability Ratings
An independent objective assessment of individual military AI capability RBD Rating (coming soon)
