This workshop uses design fiction (Bleecker et al., 2022) to explore the near future of Large Language Models (LLMs) and Generative AI (Wolfram, 2023a) as they will manifest in Computer-Aided Systemic Design. Participants learn about how AI agent systems help us understand systems in the future. The workshop aims to engage attendees in imagining the implications of Generative AI for Computer-Aided Design. It introduces two fictional software systems, one which allows AI autonomy and collaboration and the other which is hierarchical, both of which interact with humans and other agents in complex environments. The workshop explores design principles (Alexander, 1977) for creating and managing AI agents based on their goals, roles, behaviours, and interactions (Baker et al., 2020). Building on the presenter’s previous work (Cramer, 2020, 2022), the workshop combines presentations, discussions, and an interactive session to enable participants to gain insights into the human management of a community of helpful AI assistants. Participants collaborate to create their own AI agents and scenarios in a hands-on pencil-and-paper exercise (Candy, 2014). The workshop concludes with a group discussion and analysis of the experiences and learnings from the exercise, as well as the potential challenges and opportunities of Generative AI for Computer-Aided Design.
KEYWORDS: agent-based modelling, Large Language Models, ChatGPT, GPT-4, scenarios, design fiction, generative AI, computer-aided design, AI ethics, AI autonomy futures