“Go generate a bridge and jump off it”: How video pros are navigating AI (arstechnica.com)

In 2016 Hayao Miyazaki famously reacted with disgust to an early AI-generated video of a deformed human, calling it an “insult to life itself.” That rejection framed public attitudes when filmmaker PJ Accetturo released an AI-generated live‑action trailer for Princess Mononoke in October 2024: the clip drew 22 million views on X, but also torrents of abuse and death threats — from “go generate a bridge and jump off of it” to calls for violence. The episode captures a broader schism in the creative world: many artists condemn AI tools as built on stolen work and fear job displacement, while others use increasingly powerful image and video generation models to speed workflows and unlock new aesthetic possibilities. For the AI/ML community the story underscores concrete technical and policy challenges. Recent model advances have made photorealistic synthetic media easy to produce, lowering the barrier to deepfakes and creative tools alike; that raises questions about dataset provenance, copyright, and opt‑out mechanisms. It also spotlights the need for technical mitigations — watermarking, provenance tracking, robust detectors — plus clearer licensing, content moderation, and industry norms that balance innovation with artists’ rights and safety. Interviews with actors, directors and creators show many are still experimenting with AI as a tool while navigating stigma, legal uncertainty and real‑world harassment.