OpenAI moves to wire Sora straight into ChatGPT

OpenAI is preparing to wire its Sora video model directly into ChatGPT, turning the company’s flagship chatbot into an end‑to‑end text‑to‑video factory that lives inside a chat window. The move builds on earlier steps that gave ChatGPT Plus and Pro subscribers access to Sora through separate interfaces, and follows public hints from OpenAI executives that native Sora controls would land “inside ChatGPT” after the standalone rollout. TechCrunch reported those plans in early 2025, and Sora’s Wikipedia entry now explicitly notes the planned ChatGPT integration as a next phase of the product’s evolution.

Today, Sora already lets users turn prompts or images into short, cinematic clips, with the Sora 2 update adding TikTok‑style social features, video stitching and powerful “cameo” tools that can insert a person’s likeness and voice into generated scenes, according to OpenAI’s own Sora 2 blog post and coverage in TechRadar. Plugging those capabilities straight into ChatGPT collapses scripting, storyboarding and rendering into a single conversational loop: users describe a scene, have the model rewrite and expand it, then spin out draft videos without ever leaving the chat.

That convenience also magnifies long‑running fears about deepfake‑grade content. Watchdog group Public Citizen has already urged OpenAI to withdraw Sora over democracy and privacy risks, calling Sora 2 a “reckless disregard” for product safety in a letter reported by the Associated Press. Researchers warn that realistic AI video is outpacing detection systems, with recent work on deepfake detectors struggling against diffusion‑generated clips from tools like Sora, Runway and Pika, according to an arXiv study. And as Axios has noted, AI video apps are already “a scammer’s goldmine,” with Sora clips easily downloaded, shared and stripped of watermarks in the wild, fueling impersonation and fraud schemes documented by Axios.

OpenAI says it is layering content watermarking, provenance metadata, detection research and policy rules on political and sexual content into Sora’s rollout, positioning Sora 2 as both a creative tool and a safety testbed in its official post. But once Sora is a tap away inside ChatGPT—already embedded across classrooms, workplaces and creator workflows—the practical question will be whether any combination of filters and labels can keep pace with the scale and speed of everyday, chat‑driven video generation.