AI robustness auditing frameworks guided by key meta-requirements
Actionable Meta-Requirements for AI Robustness Auditing
Artificial IntelligenceSoftware Engineering
Summary
AI systems need to keep working well even when things around them change unexpectedly. The authors studied lots of existing research and talked to many experts to figure out the most important things that auditing tools for AI robustness should do. They came up with 44 clear guidelines to help people check how reliable these AI systems are. This helps organizations make or improve tools that can spot weaknesses before problems happen.
What this means in practice
- •For ai risk management teams: Use the validated meta-requirements to develop or improve AI robustness auditing methods that identify system weaknesses under changing conditions.
- •For software quality assurance teams: Assess existing AI auditing tools for completeness against the 44 meta-requirements to ensure thorough robustness evaluations.
Authors
Joshua Owotogbe, Roya Nasiri, Indika Kumara, Damian Andrew Tamburri, Willem-Jan van den Heuvel
Abstract
AI-enabled systems must remain dependable when the conditions under which they operate change (e.g., due to input perturbations, distribution shifts, or changes in external dependencies). This makes robustness auditing of such systems important for judging whether the available evidence is sufficient to support a defined robustness conclusion. However, designing an AI robustness auditing framework is challenging. A key unresolved issue is determining the requirements such a framework should support; these requirements have not yet been consolidated. We address this gap through meta-requirements: higher-order requirements specifying what a class of auditing frameworks should support. In design science research, literature and domain experts are common sources of meta-requirements; accordingly, we derive the initial set from literature and validate and refine it with practitioners. Our systematic literature review retained 74 of 2,890 records and derived 43 meta-requirements across 16 categories. We evaluated them with 128 practitioners, obtaining 5,504 ratings and 586 substantive comments. Overall, 81.20% of ratings indicated agreement or strong agreement. In particular, 41 of the 43 literature-derived meta-requirements were classified as strongly supported. Practitioner feedback refined the requirements and identified an additional requirement concerning external components and supply chain dependencies. The final set comprises 44 meta-requirements across 16 categories, of which 43 received quantitative validation. These meta-requirements can help organizations implement AI robustness auditing frameworks, assess existing frameworks for completeness, and refine them.