The Blurred Truths of Sora (www.wired.com)

OpenAI’s new app Sora lets anyone generate short, highly polished AI videos on demand — think ten-second “cameos” where users prompt digital likenesses to perform surreal or comedic scenarios. Built with infinite scroll and an emphasis on fast consumption rather than user-uploaded footage (you can’t import camera-roll media), Sora reached over a million downloads in its first week and is positioned as a creative engine for entertainment and fan fiction. OpenAI’s Sam Altman frames it as a potential “Cambrian explosion” of creativity, but also warns it could feed a reinforcement‑learning‑optimized “slop feed.” That combination of technical power and viral UX makes Sora consequential beyond novelty. Experts warn it amplifies deepfake risks, erodes authenticity, and accelerates antisocial platform dynamics by prioritizing spectacle and individualized “vision” over collective, real-world discourse. Key implications: provenance and consent are harder to enforce when likenesses are prompt-generated; addictive short-form loops magnify reinforcement of attention-optimized content; and the app’s solitary creative model limits the collaborative remix culture that made TikTok generative. Sora may thrive as a movie‑theater–style entertainment space, but it also intensifies long-standing moderation, trust, and social-cohesion challenges — underscoring the need for skepticism, provenance tools, and new norms or regulation as AI-generated visual media scale.