The field of autonomous systems is witnessing significant advancements in safety-critical scenario generation, driven by the need for robust verification and validation methodologies. Recent developments focus on overcoming the limitations of existing methods, which often struggle to balance diversity and criticality in scenario generation. A key direction in this area is the integration of dual-space guided frameworks, which coordinate scenario parameter space and agent behavior space to generate diverse and critical testing scenarios. Another notable trend is the use of adversarial generation and collaborative evolution techniques to create safety-critical scenarios for autonomous vehicles, leveraging large language models and complex traffic flows to expose unforeseen failure modes. These innovations have the potential to significantly improve the safety and reliability of autonomous systems. Noteworthy papers include: DiCriTest, which proposes a dual-space guided testing framework that improves critical scenario generation by an average of 56.23% and demonstrates greater diversity. Adversarial Generation and Collaborative Evolution of Safety-Critical Scenarios for Autonomous Vehicles presents ScenGE, a framework that generates plentiful safety-critical scenarios by reasoning novel adversarial cases and amplifying them with complex traffic flows, uncovering more severe collision cases (+31.96%) on average than state-of-the-art baselines.