Generative Enactments: Configurable Co-Futuring with Reactive AI-Generated Video Scenarios
Many consequential situations are difficult to experience directly: they are unsafe to stage, ethically impermissible, or not yet possible, yet researchers still need to study how choices shape outcomes. Pre-authored futuring methods often trade off fidelity, reactivity, or scalability. Building on AI video generation that is now fast and inexpensive enough to run inside a session, we propose \emph{Generative Enactments}: a configurable methodological framework in which participants co-speculate plausible futures through reactive, AI-generated first-person video. Participants act within a scenario and receive action-contingent consequences with fast turnaround or reactive streaming, so they can see situations unfold in response to what they do. A configuration form specifies how researchers, participants, and AI systems propose, generate, contest, and revise alternative pathways for different research purposes. We illustrate the framework through domestic robots' attention allocation (adversarial, turn-based), bystander intervention (choice-based), and trainee teacher consequence rehearsal (reactive streaming), reporting implementation and pilot-use evidence rather than outcome evaluation.