AI-Assisted Video Montage as Coordination: Design Guidelines for Platforms of Interactive Agent-based Multimodal Synthesis

Current platforms for agent-based media generation show that multimodality and multi-agent stacks can aid in the development of novel synthesis methods, namely for video signals. Open platforms and applications for AI-assisted media synthesis being proposed in scientific research focus largely on agent and model functionality with roles within higher-level tasks. Building on (Arandas et al., 2023), this article approaches AI-assisted video synthesis from a coordination standpoint and enumerates a set of guidelines for the development of interactive platforms for agent-based video synthesis, following a discussion on metacreation literature. In this workshop we will share our research as a set of guidelines for the development of new platforms for interactive agent-based video synthesis, where heterogeneous sets of models are used to produce video signals in coordination.