OpenAI Launches Sora 2: Turns Text into Video in the Blink of an Eye, Hard to Tell Real From Fake

Sora 2 Introduction Video - Video: OpenAI

Sora is a text-to-video system developed by OpenAI, first appeared in 2024. This technology allows text descriptions and receives a vivid video, creating the feeling of watching a real scene.

With Sora 2, scene-setting capabilities are taken to new heights, aiming to transform human imagination into images and sounds more realistic than ever.

Allows users' faces and voices to be added to AI videos

Sora 2 is notable for its ability to turn text descriptions into high-fidelity audio and video. Users enter a text, and the system creates a dynamic scene with synchronized characters, scenes, and dialogue.

The big improvement is that the character's mouth movements match many different languages, making the video more convincing and difficult to distinguish between real and fake.

A new feature called cameos allows users to consent to having their face or voice appear in AI videos. This opens the door to personalized content, but only works with explicit consent to limit the risk of spoofing.

Control over footage has also been improved. Creators can choose the location, duration, and transition effects, making videos more seamless and detailed.

Additionally, OpenAI applies a set of censorship rules, preventing requests to create sensitive content or use celebrity images without consent.

From journalism to education to small business: opportunities with Sora 2

Some videos created by Sora 2 - Video: OpenAI

Sora 2's fast video creation capabilities open up a wide range of real-world applications. Content creators can save time and money by creating scenes to illustrate a script. A journalist can turn a short description into a vivid illustration, or a history teacher can recreate an ancient event to help students visualize it.

Small businesses can also produce simple commercials without a studio or actors.

OpenAI is looking to integrate Sora 2 into ChatGPT and its mobile app, making it accessible to the general public. The idea of turning the tool into a social-network-like video sharing platform has also been tested, promising to bring a completely new way of creating content. Microsoft has even tried to integrate the technology into Bing, allowing short videos to be created directly on the phone.

With these steps, Sora 2 is not only a tool for professionals but also a creative assistant for everyone in many fields.

Fake video risks and privacy

However, Sora 2 still has its limitations. Some test videos show that character movements are not completely natural, with details that "jump" or do not follow the laws of physics. When shooting close-ups of faces, the lighting and expressions are sometimes not really accurate. This is a sign that artificial intelligence still cannot fully understand the physical world .

The bigger risk is that the technology can be exploited to create fake videos. Clips created from Sora 2 are so realistic that they can easily cause confusion if shared widely on social networks. This poses a challenge for both managers and the community when faced with the risk of misinformation spreading quickly.

There are also questions about privacy and copyright. If a video shows someone's face or voice, who owns the rights to that image and who is responsible if it is misused?

Sora 2 marks a significant step forward as AI-generated video becomes increasingly indistinguishable from real-life footage. In the coming years, the technology could change the way ads, demonstrations, and online lectures are produced. The ability to simply enter a short description to get a finished video is a significant cost savings and gives individuals and small businesses access to a tool that was previously the preserve of major studios.

Along with the opportunity comes the need for verification. Content-sharing platforms may need to add detection filters or clearly label AI-generated videos. Users will also gradually develop the habit of checking the source before trusting an image or video.

On the positive side, Sora 2 opens up possibilities for online education, gaming and virtual reality experiences. These are areas that require vivid images, where AI can bring speed and flexibility far beyond traditional filming techniques.

Back to topic

THANH THU

Source: https://tuoitre.vn/openai-ra-mat-sora-2-bien-van-ban-thanh-video-trong-chop-mat-kho-phan-biet-that-gia-20251002160040731.htm