ChatGPT Introduces New Sora App for AI-Created Videos

What Happened?

OpenAI’s ChatGPT has introduced a new application called Sora, which allows users to create AI-generated short video clips from text or imagery. According to OpenAI’s website, the newest Sora app, unofficially known as Sora 2, is more physically accurate, realistic, and controllable than prior systems. It also features synchronized dialogue and sound effects. 

The new feature is currently available to Apple users only and requires an invitation to access. Sora is a general-purpose video-audio generation system, which means it can create sophisticated background soundscapes, speech, and sound effects with a high degree of realism.

Why it Matters

Many AI art programs can create images from text descriptions, meaning the user only needs to type in a description of the picture they want to see and hit generate, then the AI program will do the rest and create the desired image. Sora works in a similar way, only it is an AI program that creates short videos instead of just still images. 

Sora is available to ChatGPT subscribers for a fee ranging from $20 to $200. For higher fees, users can get a more capable version of the app, which includes more videos per month for creation, higher definition videos, and other advanced features not available to basic subscribers. Presently, Sora is only available for Apple users, but that is likely to be expanded in the future.

The ability to create realistic video footage simply from a text command is potentially a double-edged sword for the evolution of AI. On the positive side, the new technology gives users with no professional training in content creation the ability to create videos that look like they came from a Hollywood studio. Ideas that might never have been turned into video content now can be, and in the modern media landscape, anything can go viral and receive mass attention.

On the negative side, there are several potential problems. One is copyright issues, with some users already complaining that the new Sora app is being used to turn out videos using trademarked images or logos without permission. Another potential problem is the use of AI-generated videos to impersonate nearly anyone, from elected officials and celebrities to ordinary people. 

Highly realistic videos showing real people doing things they never actually did are already increasing online, and the new Sora program is likely to send that number even higher. OpenAI says there are guardrails on Sora to prevent unauthorized use, but early users say the guardrails are easy to evade.

How it Affects You

While Sora could allow new artistic creations to see the light of day, it may also contribute to the amount of AI slop online, which is already growing at a rapid pace. With AI videos that are indistinguishable from real footage, already plummeting trust in online content may decrease even more.