Engadget Podcast: How to survive our AI video hellscape (www.engadget.com)

Engadget’s latest podcast dives into the uncomfortable reality that AI-generated video is no longer niche — Google and OpenAI’s new video-generation models (including OpenAI’s Sora) are producing increasingly convincing footage that’s already seeping into official channels and news cycles. Guests Drew Harwell and Jeremy Carrasco (ShowtoolsAI) discuss how these models threaten our ability to distinguish real from fake, highlighting practical harms (misinformation, political manipulation, and erased provenance) and the urgent need for media literacy, detection tools, and policy guardrails. Carrasco brings a producer’s perspective on how workflows and live production must adapt, while Harwell stresses verification, watermarking, and metadata as immediate defenses. The episode also notes a parallel hardware story: Apple’s M5 MacBook Pro and iPad Pro deliver notable graphics and compute boosts that will accelerate both generation and consumption of AI video, with a preview of the Vision Pro M5 review. For the AI/ML community this matters because improved models plus wider, faster hardware distribution lowers the barrier to high-quality video synthesis — raising research priorities around robust detection, provenance standards, efficient watermarking, and tooling that helps creators and platforms manage authenticity at scale.