OpenAI is folding Sora into ChatGPT, turning the world's most popular chatbot into a video factory.
The Signal
Sora launched as a standalone app less than a year ago and flopped. Nobody wanted another app to download, another login to remember, another subscription to manage. Meanwhile, ChatGPT already has 200+ million weekly active users who fire it up reflexively. The math is obvious: put the video generator where the people already are.
This mirrors what OpenAI did with DALL-E integration. Image generation became just another thing ChatGPT could do, buried in the flow of conversation. You're already chatting, you need a visual, boom, it's there. No context switching. No friction. That's how capabilities become habits.
But video is different from images. A deepfake photo is one frame of deception. A deepfake video is a narrative. It moves, talks, convinces. The Sora app already produced realistic clips that fooled people. Now OpenAI is about to put that power in the hands of everyone already using ChatGPT for homework help and email drafts. The barrier to creating convincing fake video just dropped to zero.
The agent economy doesn't just need tools that work. It needs tools that are already in the workflow. OpenAI knows this. Embedding Sora in ChatGPT isn't a product decision. It's a distribution strategy. And distribution, not capability, is how the AI wars get won.
The Implication
Watch for the response from Anthropic and Google. They can't let OpenAI own the integrated multimodal experience. Expect video generation in Claude and Gemini within months. For creators and communicators, the question shifts from "can I make video" to "how do I prove what's real." Provenance tools, metadata standards, verification layers. That's the next build.