Explainable AI

Motional and MIT have developed CW-Net, a system that translates self-driving car neural network logic into human-understandable concepts. Tested in real-world Las Vegas scenarios, CW-Net provides real-time explanations for vehicle decisions, aiding in debugging and building trust. This “causally faithful” approach maintains performance while significantly enhancing AI transparency, a crucial step for autonomous vehicle adoption and regulatory compliance.

DXC Technology launched Xponential, an AI orchestration framework, on October 28, 2025, to simplify enterprise AI adoption. Built on five pillars – Insight, Accelerators, Automation, Approach, and Process – Xponential addresses governance, scalability, and responsible AI. DXC cites client successes including reduced service desk tickets at Textron and improved patient care at Singapore General Hospital. DXC plans to scale Xponential globally using its 50,000 engineers and global AI centers. Analysts emphasize the importance of demonstrating ROI, integrating with existing IT, and offering explainable AI for widespread adoption.

The EU can lead in AI by leveraging its data protection regulations and open data principles. Instead of mirroring US or Chinese approaches, the EU should focus on clear data access rules, promoting accessible, high-quality datasets through initiatives like common European data spaces. Investments in federated learning, privacy-preserving techniques, and explainable AI (XAI) are crucial. This multi-faceted approach combines regulation with strategic investment, fostering responsible AI innovation and attracting investment in key areas like industrial automation and healthcare, boosting European competitiveness.