Recently, Openai introduced Sora system that can create a video, the quality of which can be compared with Hollywood films. The demonstration of the technology included short videos created in just a few minutes: mammoths walking in a snowy clearing, a monster looking at a melting candle and a street scene in Tokyo.
Among many companies seeking to improve video generation technology are both startups like Runway and Openai and the technological giants of Google and Meta*. Such a technology can speed up the work of experienced filmmakers and even replace less experienced artists. However, there are fears that video generation can become an accessible and inexpensive way to create misinformation – to distinguish reality from fiction on the Internet becomes even more difficult.
The new Openai system called Sora (from the Japanese “sky”) is currently provided to a small group of academic researchers for analysis for possible abuse. After assessing all the risks, the system will become available to a mass user. Video created by Sora are marked with watermarks that identify them as generated AI, but the company recognizes that signs can be deleted or they can be invisible.
The system is an example of generative AI that can create text, images and sounds, learning based on the analysis of digital data. Openai does not disclose, on the basis of how many videos the system studied and where they were taken, indicating only to the use of both publicly accessible and licensed video materials.
* META and its products are recognized as extremist, their activities are prohibited in the territory of the Russian Federation.