OpenAI’s latest AI can create near-perfect videos

Dall-E It is capable of creating images indistinguishable from photographs. ChatGPT reacts as a human would. TO OpenAI, The company behind these artificial intelligences needed to conquer video. Now he can also boast that he has achieved this.

The company’s newest artificial intelligence tool is littera model capable of creating high-quality video sequences. from a simple text description.

“Sora is capable of creating complex scenes with multiple characters, specific types of movements, and precise subject and background details. The model understands not only what the user has requested, but also how things work in the real world,” they explain. from OpenAI.

This isn’t the first generative engine to try to create moving scenes, but so far the results of alternatives have left much to be desired. Although tools like Dall-E. Middle of the road or Image are capable of creating very realistic static images, it is difficult for artificial intelligence to maintain consistency between various elements of the video sequence.

The results become comical, meme-like. Fingers grow on your hands, clothes change color and even size and cut. A particularly funny video circulating online shows the actor’s recreation. Will Smith there is spaghetti, in which the pasta seems to have a life of its own.

Sora has his flaws too, but he manages to create scenes that are much more realistic than the vast majority, and, perhaps more importantly, longer, down to the minute. The videos also have good resolution, 1080p equivalent. Some of the examples that OpenAI shows on its website are truly amazing and, unless you look closely at the details, are almost indistinguishable from the real video.

The engine also manages to create convincing 3D animation scenes or mix realistic objects with others that are clearly computer-generated.

He realistic results reiterates the potential of these tools in disinformation campaigns. It is relatively easy to create scenes that may appear real but are imaginary, much more than has been possible until now with advanced video and audio editing tools.

At this time, OpenAI has not opened the tool to any users. Only a small number of scientists, specialists and security specialists responsible for finding possible bugs will have access to Sora in its first versions.

The model also has the same limitations as Dall-E, a static image generation engine. You may not create videos with real personnel or with sexual or violent content. “We will engage in dialogue with policymakers, educators and artists from around the world to understand their concerns and identify positive uses for this new technology,” they explain from OpenAI.