RAID: A Framework for Embedding Responsible and Inclusive AI in Agile Software Development

As Artificial Intelligence (AI) systems become increasingly embedded in critical aspects of society, concerns about fairness, accountability, transparency, and inclusivity have become central to both academic and industry discourse. While Agile methodologies dominate modern software development, they often lack structured mechanisms to embed ethical reflection and inclusive design into day-to-day workflows. This paper introduces the RAID Framework, or Responsible and Inclusive AI in Agile Development, a conceptual model that integrates Responsible AI principles and inclusive practices directly within Agile ceremonies and artifacts. Developed using a Design Science Research methodology, RAID consists of five modular components: ethics-oriented backlog items, fairness-aware user stories, inclusive personas, a Responsible AI sprint checklist, and ethical retrospective prompts. Each component is mapped to a specific stage in the Scrum process. Scenario-based walkthroughs across various domains, including finance, healthcare, education, and public services, demonstrate the framework’s adaptability and potential to identify ethical risks without compromising team velocity. By embedding ethics and inclusion as core engineering concerns, RAID offers a scalable pathway toward building AI systems that are not only technically robust but also socially responsible.