The sector of artificial intelligence runs fast and is among the most innovative tools just presented Sora Of OpenAIan AI model that allows you to create videos starting from textual descriptions.
The peculiarity of this tool, however, lies in the fact that it is capable of creating very high resolution videos (lasting maximum of 60 seconds), creating very complex and extremely detailed scenes. Let’s find out how it works.
What Sora is and how it works
Sora allows the user to enter a more or less accurate textual description, letting the AI model make the video as precisely as possible.
To do this, it uses complex neural networks that allow it to be created consistent footageshowing a large number of details, from emotions of the protagonists and until arriving at the correct understanding of laws of physicsso that everything that appears on the screen interacts and fits into the created world in a coherent way.
In this sense, it is possible to ask subjects in scenes to move within the context naturallywith the possibility of interact with the environment very convincingly and as close to reality as possible.
This is possible thanks to a deep understanding of language by this AI model that manages to interpret (almost) every request, reproducing it in the best possible way.
Besides, Sora doesn’t even need to frame-by-frame directionsbut starting from the overall description, it is able to maintain a certain coherence in the video shot, adapting the movements automatically.
Clearly the tool it’s not perfect (yet). and in some cases there may be difficulties in understanding the requests, reproducing “glitches” and anomalous situations, even if to be honest the system evolves in an incredible way, becoming more and more efficient.
Finally, according to the statements of OpenAISora is also capable of making a video starting from a fixed image o complete existing films with other frames created ad hoc.
What are the risks for users
As stated on the official OpenAI website, for now Sora is only available for one small group of researchers who are evaluating all the relevant risks.
However, it may still take some time before it becomes available to everyone (or at least to subscription users) but it still represents a revolution in digital content creationwhich could radically disrupt the entire sector.
There are, however, several doubts about the use of this tool which could be used for creation of deepfakes extremely credible. An eventuality that cannot be overlooked and which must be addressed promptly.
For this reason, OpenAI and other companies are working on a watermark that will act as a company digitalwhich will allow users to identify the contents created with such tools and, if used incorrectly, help them avoid falling into the trap of digital attackers.
To know more:
- Artificial Intelligence: what it is and what it can do for us
- ChatGPT, what it is, how it works, what it is for, how to use it for free