Generative AI is rapidly transforming the intellectual and operational infrastructure of software architecture research. From co-generating manuscripts, models, and diagrams to augmenting literature reviews and supporting reviewers with automated summarisation and analysis, AI is reshaping how research is produced, evaluated, and disseminated. Emerging practices, such as agent-based artifact evaluation, further extend this transformation. Collectively, these shifts impact not only productivity, but also the rigor, reproducibility, and integrity of architectural research. As a result, academic conferences themselves are evolving into AI-enabled socio-technical systems.

Despite routinely architecting complex software systems, we seldom apply the same architectural thinking to the design of conference ecosystems. Conferences such as ICSA and ECSA can be viewed as long-running platforms composed of interacting components, including submission pipelines, review workflows, artifact evaluation processes, program committees, governance structures, and community norms. These components embody critical quality attributes such as fairness, transparency, traceability, scalability, inclusiveness, and sustainability. The introduction of AI is not merely a tooling enhancement, but an architectural shift that fundamentally affects these attributes.

This working session invites the community to proactively architect the evolution of AI-enabled conferences. Framed as a collaborative design exercise, we will explore key questions: Should AI be treated as a first-class architectural component in research workflows? How can governance mechanisms ensure human accountability while enabling effective augmentation? How do we design review systems that preserve trust, mitigate automation bias, and elevate human expertise? What enduring architectural principles should guide the future of AI-enabled conference systems?