Generative AI is reshaping software work, yet we lack clear guidance on where developers most need and want support, and how to design it responsibly. We report a large-scale, mixed-methods study of N=860 developers that examines where, why, and how they seek or limit AI help, providing the first task-aware, empirically validated mapping from developers' perceptions of their tasks to AI adoption patterns and responsible AI priorities. Using cognitive appraisal theory, we show that task evaluations predict openness to and use of AI, revealing distinct patterns: strong current use and a desire for improvement in core work (e.g., coding, testing); high demand to reduce toil (e.g., documentation, operations); and clear limits for identity- and relationship-centric work (e.g., mentoring). Priorities for responsible AI support vary by context: reliability and security for systems-facing tasks; transparency, alignment, and steerability to maintain control; and fairness and inclusiveness for human-facing work. Our results offer concrete, contextual guidance for delivering AI where it matters to developers and their work.

翻译：生成式AI正在重塑软件工作，然而我们对于开发者最需要且最希望获得支持的具体领域，以及如何负责任地设计此类支持，仍缺乏明确指导。我们报告了一项针对N=860名开发者的大规模混合方法研究，考察了他们在何处、为何以及如何寻求或限制AI帮助，首次提供了从开发者对其任务的认知到AI采用模式及负责任AI优先事项的任务感知型、经验验证的映射。运用认知评价理论，我们证明任务评估能够预测对AI的开放态度及使用行为，并揭示了三种显著模式：核心工作（如编码、测试）中当前使用强度高且存在改进需求；对减少繁琐劳动（如文档编写、运维）存在强烈需求；以及对身份认同与关系导向型工作（如指导）存在明确的使用限制。负责任AI支持的优先事项因情境而异：面向系统的任务需注重可靠性与安全性；为保持控制需强调透明度、对齐性与可操控性；面向人际的工作则需关注公平性与包容性。我们的研究结果为在开发者及其工作真正需要的领域提供AI支持，提供了具体且情境化的指导。