Trustworthy AI Suffers from Invariance Conflicts and Causality is The Solution

2026-05-04

AI Summary

This work addresses the inherent tensions among key desiderata of trustworthy artificial intelligence—such as fairness, robustness, privacy, and interpretability—particularly when preserving model utility. It is the first to systematically frame these multi-objective trade-offs as incompatibilities among invariances under distinct transformations in the data-generating process. The paper proposes causal inference as a unifying framework, leveraging selective invariance modeling to uncover and reconcile such conflicts. By integrating invariance learning with an analysis of data-generation mechanisms, the approach applies seamlessly to both conventional models and large foundation models. This not only offers a theoretically coherent perspective on trustworthy AI but also charts a path toward building efficient, multi-objective-compatible AI systems, while outlining critical challenges and opportunities for future research.

Abstract

As artificial intelligence (AI), including machine learning (ML) models and foundation models (FMs), is increasingly deployed in high-stakes domains, ensuring their trustworthiness has become a central challenge. However, the core trustworthy AI objectives, such as fairness, robustness, privacy, and explainability, are hard to achieve simultaneously, especially while preserving utility. This position paper argues that causality is necessary to understand and balance trade-offs in performance and multiple objectives of trustworthy AI. We ground our arguments in re-interpreting trustworthy AI trade-offs as incompatible invariance requirements under different changes to the data-generating process. We then illustrate that causality provides a unifying framework for understanding how trade-offs in trustworthy AI arise, and how they can be softened or resolved through selective invariance. This perspective applies to both classical ML models and large-scale FMs. Our paper discusses how causal assumptions may be applied explicitly or implicitly in modern large-scale systems. Finally, we outline open challenges and opportunities for using causality to build more trustworthy AI.

Problem

• Trustworthy AI
• Invariance Conflicts
• Causality
• Fairness
• Robustness

Innovation

• causality
• invariance conflicts
• trustworthy AI
• selective invariance
• foundation models