- Artificial Intelligence Safety and Governance: Theories and Methods (人工智能安全与治理理论与方法)
- Safety Risks and Protection Mechanisms in Large Models and Generative Artificial Intelligence (大模型与生成式人工智能中的安全风险与防护机制)
- Robustness, Reliability, and Defense Against Adversarial Attacks in Artificial Intelligence Models (人工智能模型鲁棒性、可靠性与对抗攻击防御)
- Trustworthy Artificial Intelligence, Explainability, and Auditable Governance Mechanisms (可信人工智能、可解释性与可审计治理机制)
- Artificial Intelligence Alignment, Controllable Generation, and Human Oversight Mechanisms (人工智能对齐、可控生成与人类监督机制)
- Safety Boundaries, Tool Use, and Behavioral Constraints in Agentic Systems (智能体系统中的安全边界、工具调用与行为约束)
- Collaborative Safety and Governance of Emergent Risks in Multi-Agent Systems (多智能体系统中的协同安全与涌现风险治理)
- Safety Control and Accountability Governance in Embodied Intelligence and Autonomous Systems (具身智能与自主系统中的安全控制与责任治理)
- Model Misuse, Deepfakes, and Content Safety Governance in Artificial Intelligence Systems (人工智能系统中的模型滥用、深度伪造与内容安全治理)
- Risk Identification, Assessment, and Tiered Governance for Artificial Intelligence (面向人工智能的风险识别、评估与分级治理)
|
- Next-generation artificial intelligence is rapidly evolving from an algorithmic tool into a complex intelligent system with capabilities for perception, generation, reasoning, interaction, and action. It is being widely applied in critical domains such as transportation, manufacturing, healthcare, finance, education, public safety, and social governance. While AI technologies are driving industrial transformation and social innovation, they also bring a range of safety and governance challenges, including model hallucinations, algorithmic bias, adversarial attacks, model misuse, unauthorized actions by agents, physical risks associated with embodied intelligence, failures in human-AI collaboration, unclear attribution of responsibility, and lagging governance mechanisms.
This forum focuses on issues of AI safety and governance. Centered on model safety risks, system reliability, agent controllability, risk assessment methods, safety evaluation frameworks, trustworthy governance mechanisms, and key enabling technologies across the full AI lifecycle, it will explore relevant theoretical foundations, methodological innovations, system implementations, standards and norms, and representative applications. The forum aims to provide a high-level platform for exchange among experts, scholars, and industry researchers in related fields, promote the development of an AI safety governance system suited to the future intelligent society, and support the transition of AI technologies from capability breakthroughs to trustworthy deployment, and from isolated applications to responsible innovation.
|
