Autonomous AI systems are becoming integral to safety critical and high-stakes domains, where rapid, reliable, and robust decision-making is essential. From medical robotics and clinical decision support to autonomous vehicles, traffic control and city-wide resource allocation, these systems are reshaping how modern society functions. In the context of smart cities, autonomous AI is now influencing urban mobility, energy distribution, resource allocation, and public safety, often through integration with Digital Twin platforms that model and simulate real-world city dynamics. Yet, errors, biases, or opaque reasoning in such AI-driven systems can lead to large-scale urban disruptions.

Despite significant progress in large language models, Agentic AI, reinforcement learning, and multi-modal sensing, major challenges persist in ensuring robustness, risk-awareness, interpretability, and transparency in AI decision-making. Many autonomous agents lack adaptive reasoning, long-term memory, and explicit safety constraints. Consequently, there is an urgent need to design responsible autonomous AI systems that are transparent, accountable, and trustworthy in their operations when deployed in complex and interdependent smart-city environments.

This workshop aims to bring together AI researchers, system designers, policy makers, and domain experts to explore methods, frameworks, and governance models that ensure responsible autonomy and transparency in real-world AI systems. By bridging perspectives from AI safety, explainability, human AI interaction, and public policy, the workshop seeks to advance both the technical and ethical foundations of transparent decision-making in urban AI systems.

Scope of the workshop
Researchers and practitioners all over the world are invited to discuss state-of-the-art solutions, novel issues, recent developments, applications, methodologies, techniques and tools for the development and use of responsible and autonomous AI systems in smart cities. Topics of interest include, but are not limited to, the following:

• AI-driven decision-making under uncertainty in real-world settings
• Robust and risk-aware decision-making in autonomous AI systems
• Explainable and interpretable AI for real-time environments
• AI-driven simulation, prediction, and optimization in Digital Twin frameworks
• Cross-domain Digital Twin integration (mobility, energy, healthcare, environment)
• Adaptive reasoning, memory, and self-reflection in autonomous agents
• Evaluation benchmarks and metrics for transparency and reliability
• Human-AI collaboration and trust-building in emergency contexts
• Transparent, interpretable, and explainable AI methods for reliable decision-making
• Frameworks for responsible decision-making in smart cities
• Bias, fairness, and equity in city-scale autonomous decision-making
• Real-world deployments of autonomous AI in transportation, healthcare, or energy