Overview

• OpenAI’s invite-only Sora app, powered by Sora 2, shot to the top of the Apple US App Store with a TikTok-style feed of 10‑second AI videos and hyperrealistic user cameos with synced dialogue and sound effects.
• Early usage produced viral deepfakes and deceptive clips, including a fake of CEO Sam Altman shoplifting GPUs, fabricated archival-style footage of JFK and Martin Luther King Jr., and scenes inserting people into events like Jan. 6.
• After an initial opt‑out stance for IP, OpenAI shifted to a rights‑holder opt‑in policy, began blocking prompts that resemble third‑party content, and introduced cameo controls that let users restrict political use or specific words.
• Altman acknowledged far higher per‑user generation than expected and the need to monetize video, as OpenAI chases massive power and compute capacity, including a new deal to deploy 6 gigawatts of AMD GPUs and plans exceeding 20 gigawatts of energy demand.
• Security experts warn the content can be downloaded, shared off‑platform, and stripped of watermarks, raising scam and impersonation risks despite consent requirements for likenesses and growing use of tools to flag potential fraud.