Explainable Artificial Intelligence (XAI) for 6G: Improving Trust between Human and Machine

Trust requires transparency as AI becomes more autonomous: The paper argues that as 6G networks move from traditional model-based systems to deep learning approaches, "the trust we have in our optimisation modules decrease" because we lose the ability to understand why the AI makes specific decisions—a critical gap for mission-critical services like autonomous driving and remote surgery.

Explainability is both a technical and legal necessity: The paper identifies "public and legal motivations" for XAI in 6G, indicating that regulatory requirements and public accountability demands are driving the need for explainable systems, not just technical preferences.

There's a real trade-off between AI performance and explainability: The paper specifically highlights "performance vs. explainability trade-offs" as a core concept, suggesting that making AI systems more understandable may require accepting some reduction in their optimization capabilities—a practical constraint decision-makers need to understand.

Current AI systems cannot account for data quality problems: The paper emphasizes that without explainability, engineers and the public cannot understand "the impact of: 1) poor/bias/malicious data, and 2) neural network design on decisions," meaning hidden biases or corrupted data could influence critical network decisions without detection.

XAI methods and frameworks are being developed for wireless networks: The paper outlines that the field is moving beyond theory by providing "methods to improve explainability, and frameworks to incorporate XAI into future wireless systems," suggesting practical tools are emerging to address these trust challenges.