OpenAI, the artificial intelligence company responsible for ChatGPT, launched a new model called Sora, which transforms text into videos of up to one minute in a “realistic” and “creative” way. The new feature was announced on Wednesday (14) and is in the testing phase, being made available to a select group of users.
According to OpenAI, the tool is available for “red teamers“, a group of professionals specializing in cyber threats, to assess risks and ensure that artificial intelligence does not produce harmful or inappropriate content. In addition, the company is also granting access to visual artists, designers and filmmakers to get feedback on how to make the model more useful for these professionals.
“We are sharing our research progress early to get work done and get feedback from people outside of OpenAI and to give the public a sense of what AI capabilities are on the horizon,” the company wrote in a statement. announcement.
What is Sora and how will the model work?
According to OpenAI, Sora is capable of creating videos of up to 60 seconds from text instructions, with complex scenes with characters, specific types of movements and precise details. “The model understands not only what the user asked for in the prompt, but also how these things exist in the physical world,” says the tool’s announcement.
To generate these images, Sora has a “deep understanding of language,” which allows him to accurately interpret instructions and generate compelling characters who are capable of expressing emotions. You can check out videos created with AI here.
On the other hand, the company emphasizes that Sora is still a work in progress with “weak points” and may have difficulty accurately simulating details of depth and space, potentially confusing left and right and understanding cause and effect situations. For example: a person might ask the model to create a video of someone taking a bite of a cookie, but the cookie might not have a bite mark.
Additionally, OpenAI states that it will take several security measures before making Sora available in the company's products. In addition to the tests carried out by red teamerstools are also being built to detect fake content and warn that the video was generated by Sora.
Another security measure announced by the company is the text classifier that will check and reject requests that violate usage policies, such as incitement to violence, sexual content, hateful images, celebrity likenesses or third-party IP. Image classifiers were also developed to review the frames of each generated video to ensure compliance with usage policies before being published.
“We will engage policymakers, educators and artists from around the world to understand their concerns and identify positive use cases for this new technology,” says the company.
Source: CNN Brasil
Charles Grill is a tech-savvy writer with over 3 years of experience in the field. He writes on a variety of technology-related topics and has a strong focus on the latest advancements in the industry. He is connected with several online news websites and is currently contributing to a technology-focused platform.
