The next revolution in the world of artificial intelligence.  Sora creates recordings based on text alone

The next revolution in the world of artificial intelligence. Sora creates recordings based on text alone

OpenAI presented a new technology that can once again revolutionize the use of artificial intelligence in everyday life. Sora is capable of creating one-minute recordings based on descriptions alone. OpenAI’s latest achievement impressed many Internet users. However, there are also voices pointing to the threats that may arise from technology.

Sora is a new artificial intelligence model developed by OpenAI, the organization responsible for, among others, ChatGPT and DALL-E. While the last two allow you to generate text responses and computer graphics, Sora is expected to open completely new perspectives. The latest artificial intelligence model was developed to generate 60-second recordings.

OpenAI presents a new model of artificial intelligence. Sora can create a video based on the description

After announcing the progress so far, OpenAI published a series of posts on social media showing various fragments of recordings generated by Sora. Their content also includes hints presented to the model, often referred to as “prompts”. The demonstration of the new model’s abilities allows us to determine what effects the introduction of a specific description can cause.

On the OpenAI website we read that Sora can, among other things, generate complex scenes with multiple characters, specific types of movement and precisely reproduced details of objects and backgrounds. “The model understands not only what the user has asked for, but also how these things exist and behave in the physical world,” the organization describes. Sora is also capable of reflecting various visual styles – it can, among other things, recreate a video that looks like an animation, as well as one that resembles a live-action film. It can also create multiple takes within one generated recording.

Commentators point to possible threats. The model is still in the testing phase

The presentation of the new OpenAI model received wide coverage on social media and surprised many Internet users. However, among the comments, it is also easy to find critical voices that point to possible threats from similar technologies. The organization itself emphasizes that it is aware of the dangers and at the moment it is making Sora available only to a group of testers who are to assess it in terms of possible harm and threats. According to the announcements, in the future the model will reject, among others, content that it considers to be brutal violence or sexual content, as well as materials containing similarities to publicly known people and objects constituting someone’s intellectual property.

Source: Gazeta

You may also like

Immediate Access Pro