System engineering involves a delicate interplay between three tasks: specification (what a system should do); implementation (what a system actually does); and verification (determining whether they agree). Already, novel generative AI technologies have emerged that can assist with implementation and verification (e.g., respectively, Microsoft’s CoPilot, and formal verification tools developed on DARPA’s PEARLS Artificial Intelligence Exploration). Yet the first task, specification, remains a largely manual and time-consuming process. This begs the question: Could a system leveraging generative technologies assist with specification? If so, such a breakthrough could represent a paradigm shift for how systems are designed, implemented, and maintained. AI assistants could dramatically reduce the human time and effort required to develop specifications, enabling rapid prototyping, automating more rote elements of the process to allow engineers to focus on more complex tasks, and ultimately reducing costs as the number of engineer hours needed drops. Done well, AI assistance could also help mitigate human error, improve specification quality, and improve consistency – all leading to system verification processes that are simultaneously more rigorous and more efficient. Recent advances in AI raise the possibility that generative technologies could plausibly assist operators in the authoring of specifications which meet the necessary requirements, constraints, and challenges. That is, AI systems may be able to help designers generate system specifications that are accurate, complete, and sufficiently granular to support system implementation and verification of corresponding code in a way that also enables continuous integration, deployment, verification, and certification: a kind of AI-assisted DevSecOps. To define some terms, by a specification, we mean a clear description of desired behavior and structure for some system component. This definition is intentionally...