OpenAI Integrates Sora into ChatGPT: A Risky Bet Against Deepfakes

OpenAI Expands ChatGPT with Sora

OpenAI is considering integrating its video generator, Sora, directly into ChatGPT, according to information from The Information. Until now, Sora was only accessible through its own website or as a standalone app, having not achieved the popularity of ChatGPT. This integration would allow users to create videos directly within ChatGPT, similar to the previous addition of image generation.

Increased Risks of Deepfakes

The integration of Sora into ChatGPT could attract more users, but it also raises concerns about the rise of deepfakes. At its initial launch, Sora was used to create realistic and controversial deepfakes, including videos of historical figures like Martin Luther King Jr. and copyrighted content.

Accessibility and Bypassing Protections

With Sora integrated into ChatGPT, its accessibility would be greatly enhanced, potentially leading to an increase in deepfakes. Users might also seek to bypass OpenAI's protections, such as removing the watermark that identifies AI-generated videos. Historically, users have often found creative ways to deceive AI generators.

Competition and OpenAI's Strategy

This initiative comes as OpenAI faces increasing competition. Anthropic's Claude model has recently gained popularity, while ChatGPT has seen a rise in uninstalls. Anthropic declined a contract from the Pentagon to use Claude for military purposes, which has bolstered its public support. In contrast, OpenAI accepted these terms, leading some users to turn away from ChatGPT. The integration of Sora could be a strategy to win back these users.

Financial Implications for OpenAI

However, as noted by The Information, the integration of Sora could also increase costs for OpenAI. This might lead to adjustments in ChatGPT's pricing, which has already begun displaying ads for its lowest-paying users last month.