Contents
Video from text: The technology is not new, but it has never been as sophisticated as in the new OpenAI software “Sora”.
Hardly any other technology currently gets people as excited as the achievements of artificial intelligence (AI). The last trick of the company OpenAI, which gained fame thanks to the chatbot ChatGPT, is this Text-to-video AI software «Sora». We tell you what you need to know. And whether you need to be afraid.
The software: The AI model called “Sora” will initially be made available to selected creative people, writes OpenAI boss Sam Altman on X. Experts should also explore possible security risks before the program can be used widely. In this version, amazingly consistent videos can be created with just a few textual specifications. Videos created by Sora can be up to a minute long.
On the software website, OpenAI published several examples along with the description on which they were based. One of them shows a woman walking on a street. The video was generated entirely by artificial intelligence, with the text specifying that the woman should wear a leather jacket and a red dress and that the street should be reminiscent of Tokyo and have lots of neon signs that were also reflected in puddles. For the head of SRF-Digital, Guido Berger, “Sora” is “nothing surprising” from an expert perspective. The technology built into Sora has been in the works for a long time.
The quality of the software: According to Berger, the results of “Sora” are still “magical” and look very good. For example, it is impressive how smoothly the software can combine two different videos into a new clip. In contrast to other text-to-video AIs, OpenAI is able to design the individual images within a video in such a way that moving objects hardly change from image to image. This is what distinguishes “Sora” from previous models.
However, the underlying technology is also known to other companies, so Berger expects that Google, for example, will soon be on the same level as “Sora”. However, OpenAI itself admits that the program still has certain teething problems.
The model sometimes makes mistakes when implementing physics rules. It could also happen, for example, that someone in the video takes a bite of a cookie – and the cookie later still looks whole.
The opportunities and risks of “Sora”: AI technology that generates moving images from text will transform video production over time. At the same time, there are great concerns that fake videos could be created on a large scale that can hardly be distinguished from real recordings. The developers of the technology are therefore working on ways to incorporate unique identifying features such as watermarks into the videos. The company writes that Sora videos should also show that they were created by AI.
SRF digital director Guido Berger doesn’t believe in excessive alarmism. Even before OpenAI, photos and videos could not be fully trusted. In this respect, “Sora” is nothing qualitatively new, but above all quantitatively. You can now artificially produce much larger amounts of video material, and that could put fact-checkers above the masses. On the other hand, the technology also has many advantages, says Berger. With “Sora” it was possible to create cinematic worlds that previously required expensive studio technology.