AI 中文总结
针对AI时代软件与系统工程的转型需求，本文提出ACCEL框架，构建智能体工程师的教育架构，明确核心能力与实施路径，识别风险并主张开展系统性教育变革。
AI 中文摘要
生成式与智能体人工智能（AI）正在重构软件与系统工程，使其从以人工制品的人类制作为核心的学科，转向聚焦于指挥、验证与管控自主系统的学科。这一转变催生了新的职业原型——智能体工程师，其持久价值在于意图规范、多智能体工作流编排、对机器生成输出的批判性评估以及伦理判断。本文整合了工程教育、计算机教育、人机交互、人为因素及学习科学领域的研究，构建了基于证据的该职业原型教育架构。我们提出了ACCEL框架（Agentic Competencies through Curricula, Collaboration, and Enduring Learning），该框架整合了五大能力支柱，并将其映射到课程、协作与持续学习三大实施路径。基于智能体理论、自动化信任研究及AI辅助编程的实证研究（包括AI效益分布不均且常被误解的证据），我们提出了分阶段课程、人机协作的委托-验证教学循环、重新设计的评估、具备治理素养的伦理整合，以及与现行课程指南和国际AI能力框架的对齐方案。我们还识别出关键风险，包括自动化偏差、技能退化、浅层参与及责任分散，并得出结论：培养智能体工程师需要系统性变革，而非渐进式课程调整——教学必须从人工制品生产转向对日益自主的社会技术系统行使判断权。
英文摘要
Generative and agentic artificial intelligence (AI) are reconfiguring software and systems engineering from a discipline centered on human authorship of artifacts to one focused on directing, verifying, and governing autonomous systems. This transition demands a new professional archetype, the \emph{agentic engineer}, whose enduring value lies in intent specification, orchestration of multi-agent workflows, critical evaluation of machine-generated outputs, and ethical judgment. This article presents an integrative conceptual synthesis across engineering education, computing education, human--AI interaction, human factors, and the learning sciences to derive an evidence-grounded educational architecture for this archetype. We introduce the ACCEL framework (Agentic Competencies through Curricula, Collaboration, and Enduring Learning), which organizes five competency pillars and maps them to three delivery vectors: curricula, collaboration, and continuous learning. Drawing on agency theory, trust-in-automation research, and empirical studies of AI-assisted programming, including evidence that AI benefits are unevenly realized and often misperceived, we propose a scaffolded curriculum, a delegation--verification pedagogical loop for human--AI teaming, redesigned assessment, governance-literate ethics integration, and alignment with current curricular guidelines and international AI competency frameworks. We identify key risks, including automation bias, deskilling, superficial engagement, and diffuse accountability, and conclude that educating the agentic engineer requires systemic transformation rather than incremental curricular change: instruction must shift from producing artifacts to exercising judgment over increasingly autonomous socio-technical systems.
