OpenAI unveiled its new AI video model, Sora 2, on Sept. 30, with capabilities to produce hyper-realistic videos from a simple prompt written by the user.

These videos are so realistic that it is often difficult for viewers to even recognize that it is an artificially-produced video. The topics of the video can range from reinvented videos of Bob Ross to a video of a hermit crab with a lightbulb shell.

Sora 2 by OpenAI is an app currently available for download on iOS, but it is also available for use on a web browser.

It can be a fun, lighthearted tool used to create any video imaginable, such as a video of Adam Sandler petting a leopard or Jimmy Kimmel riding a dinosaur. Unfortunately, many have used these AI capabilities to spread misinformation according to The New York Times. 

Fortunately, Sora’s watermark helps alert viewers that the content they are viewing is AI generated, but if users don’t recognize the watermark they may still believe the content is real.

Users can upload their own photos and videos and instruct the AI model to alter it however they please, and the videos take only about a minute to generate, after which they then can be downloaded and shared.

If the AI model detects a face being uploaded they require a video of the person saying a given three numbers as a form of consent.