Empirically validate the AI-ER aggregation rules across industries and business models

Determine empirically whether the maximum rule for AI exposure and the minimum rule for AI resilience are suitable across different industries and software-based business models.

Background

The AI-ER framework aggregates exposure using the maximum of substitutability and replicability, on the assumption that one pronounced attack path can determine a company’s exposure. It aggregates resilience using the minimum of protective positions and adaptability, on the assumption that one weak core condition can constrain resilience. These aggregation choices are conceptual rather than empirically calibrated.

The paper identifies the generalizability of these rules as unresolved across industries and business models. Resolving the issue requires empirical testing of whether the aggregation logic produces valid and useful classifications in settings with different forms of value creation, competitive structure, and organizational capability.

References

The current scale anchors, thresholds, combination rules, evidence weights, and modulators have not been calibrated on a large sample. The maximum rule for exposure and the minimum rule for resilience are conceptual choices. They assume that one pronounced attack path can determine exposure and that one weak core condition can limit resilience. Their suitability across industries and business models remains an empirical question.

— AI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business Model  (2609.11321 - Mandl et al., 10 Sep 2026) in Section 7.2, “Limits of the Assessment Framework” (Section 7, subsection “Limits of the Assessment Framework”)

The proposed rating logic makes AI-ER operational, but the framework has not yet been validated empirically.

— AI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business Model  (2609.11321 - Mandl et al., 10 Sep 2026) in Section 7, “Validation and Limits,” especially Section 7.1, “Approach to Empirical Validation”

Conformal validity is verified marginally and by rolling origins, but not conditionally on regime: a band can be narrow and correct on average while failing in the regime that matters. Every threshold above is calibrated against the generator and none against a commercial category. And promotional incrementality is outside the system's scope: promotion is treated as a confounder to be removed, never as a lever to be optimized.

— ACT, WAIT, or EXPERIMENT: A Causal Governance Framework for Retail Price Optimization Under Abstentions  (2609.10615 - Cadahia, 8 Sep 2026) in Appendix, Section “Declared open gaps” under Governance Thresholds and Correspondence with the Implementation