Call for Papers 征稿启事：负责任的人工智能：创新与伦理的交汇

Sustainable AI has become a central concern as artificial intelligence systems expand in scale, autonomy, and societal impact.

随着人工智能系统规模、自主性及社会影响的持续扩大，可持续人工智能已成为一个核心议题。

AI is increasingly positioned both as a tool for addressing environmental and societal challenges and as a technology with significant ecological footprints. The concept extends Green AI by integrating environmental efficiency with social, economic, and organizational sustainability considerations. For AI researchers and engineers, sustainability raises foundational questions about design choices, architectural decisions, and system-level trade-offs. Sustainability should therefore be treated as a core quality attribute of AI systems, alongside performance, reliability, usability, security, and energy-aware software behavior.

人工智能日益被定位为应对环境与社会挑战的工具，同时其本身也具有显著的生态足迹。可持续人工智能的概念通过将环境效率与社会、经济和组织的可持续性考量相结合，拓展了绿色人工智能的内涵。对于人工智能研究人员和工程师而言，可持续性引发了关于设计选择、架构决策和系统层面权衡的基础性问题。因此，可持续性应与性能、可靠性、可用性、安全性及节能软件行为一道，被视作人工智能系统的核心质量属性。

Current AI development practices prioritize accuracy and scale, often neglecting energy consumption, lifecycle costs, and long-term maintainability. The computational intensity of modern AI models places growing pressure on energy infrastructures, data centers, and hardware supply chains. Increasing model complexity challenges the feasibility of deploying AI systems on consumer-level hardware without undermining user expectations. Engineering sustainable AI requires empirical evidence on trade-offs between performance, resource usage, deployment constraints, software architectures, and system evolution over time. Technical challenges intersect with ethical concerns related to data practices, bias, accountability, security, and workforce transformation driven by automation. These challenges are further amplified by regulatory uncertainty, limited transparency of environmental impacts, and uneven distribution of AI costs and benefits across societies.

当前人工智能开发实践以准确性和规模为首要目标，往往忽视能耗、生命周期成本和长期可维护性。现代人工智能模型的计算强度给能源基础设施、数据中心和硬件供应链带来日益沉重的压力。不断增加的模型复杂性挑战了在不影响用户体验的前提下，在消费级硬件上部署人工智能系统的可行性。可持续人工智能的工程实现需要有关性能、资源使用、部署约束、软件架构及系统随时间演进之间权衡的经验证据。技术挑战与涉及数据实践、偏见、问责制、安全及自动化驱动下的劳动力转型等伦理关切相互交织。监管不确定性、环境影响透明度不足以及人工智能成本与收益在社会各阶层的不均衡分布，进一步放大了这些挑战。

This Theme Collection aims to advance sustainable AI at the intersection of engineering innovation and ethical responsibility. It seeks contributions proposing methods, measurement techniques, metrics, tools, and empirical studies supporting sustainability throughout the AI system lifecycle. The collection welcomes interdisciplinary work connecting technical evidence with ethical analysis, governance frameworks, policy design, and organizational decision-making. It also encourages reflection on how AI systems shape environmental outcomes, social equity, and access to innovation. The goal is to foster evidence-based dialogue enabling responsible, efficient, and socially legitimate AI innovation aligned with long-term environmental and societal values.

本专题旨在推动工程创新与伦理责任交汇点的可持续人工智能研究。我们诚邀各方投稿，提出支持人工智能系统全生命周期可持续性的方法、测量方法、指标、工具及实证研究。本专题欢迎连接技术证据与伦理分析、治理框架、政策设计及组织决策的跨学科研究。我们也鼓励对人工智能系统如何影响环境结果、社会公平及创新可及性的反思。目标是促成基于证据的对话，使人工智能创新得以负责任、高效率且符合社会合法性，与长期环境和社会价值保持一致。

Topics for this call for papers include but are not restricted to:

本期征稿议题包括但不限于：

• Energy-efficient and secure AI techniques, including model compression, efficient training, neuro-symbolic approaches, and empirically validated trade-offs between performance, robustness, scale, and resource consumption.

高能效且安全的人工智能技术，包括模型压缩、高效训练、神经符号方法，以及性能、鲁棒性、规模与资源消耗之间经过实证验证的权衡

• Energy-aware software architectures for AI, including web-based systems, edge–cloud trade-offs, green DevOps/MLOps, and programming abstractions supporting responsible and sustainable deployment.

面向人工智能的能耗感知软件架构，包括基于Web的系统、边缘-云端权衡、绿色开发运维/机器学习运维，以及支持负责任和可持续部署的编程抽象

• Measurement, transparency, and empirical methods for AI sustainability, including energy consumption estimation, carbon accounting, lifecycle assessment, benchmarking, reproducible evaluation, and responsible impact reporting.

人工智能可持续性的测量、透明度与实证方法，包括能耗估算、碳核算、生命周期评估、基准测试、可复现评估及负责任的影响报告

• Empirical software engineering for AI systems, addressing maintainability, technical debt, continuous learning, system evolution, and long-term responsibility of AI-enabled software.

面向人工智能系统的实证软件工程，涵盖可维护性、技术债务、持续学习、系统演进及人工智能赋能软件的长期责任

• Data and model responsibility, including responsible data practices, bias mitigation and unbiasing techniques, data reuse, and the environmental and social costs of large-scale datasets.

数据与模型责任，包括负责任的数据实践、偏见缓解与去偏技术、数据复用，以及大规模数据集的环境与社会成本

• Socio-technical and ethical dimensions of responsible AI, including accountability, fairness, trust, social justice implications, workforce transformation, and societal impacts of AI adoption.

负责任人工智能的社会技术与伦理维度，包括问责制、公平性、信任、社会正义影响、劳动力转型及人工智能应用的社会影响

• Governance, policy, and regulation for responsible AI, including standards, organizational policies, public-sector deployment, evidence-based regulation, and international or cross-border governance frameworks.

负责任人工智能的治理、政策与监管，包括标准、组织政策、公共部门部署、基于证据的监管及国际或跨境治理框架

• Industrial, public-sector, and climate-related AI applications, reporting empirical evidence, lessons learned, benefits and limits of AI for sustainability, and real-world deployment constraints.

工业、公共部门及气候相关的人工智能应用，报告经验证据、经验教训、人工智能在可持续领域的效益与局限，以及实际部署约束

Prof. Roberto Vergallo

Pegaso University

Italy

罗伯托·韦尔加洛 教授

意大利 帕加索大学

Prof. Maja Hanne Kirkeby

Roskilde Universitet

 Denmark

 玛雅·汉娜·基尔克比 教授

丹麦 罗斯基勒大学

 Prof. Ana Oprescu

Universiteit van Amsterdam

 the Netherlands

 安娜·奥普雷斯库 教授

  荷兰 阿姆斯特丹大学