The risk of AI getting out of control

The development of AI brings both positive and negative aspects. And with the appearance of Sora, the technology will become hotter than ever.

Sora - OpenAI's "new sky"

Sora is a huge step forward for OpenAI as it allows for text-to-image conversion with incredible quality.

In Japanese, “Sora” means sky. The team behind the technology chose the name Sora because it evokes the idea of limitless creative potential.

Sora is capable of creating animations with multiple characters, moving with detailed backgrounds. The tool not only understands the suggestions in the command but also how it manifests in reality. According to reviews, Sora has the ability to speed up the work of filmmakers, at the same time completely replacing less experienced graphic engineers.

Sora is not the only AI program that has successfully converted text into images. Over the past year, there have been many other programs that have the same function. The products created by AI technology are becoming more and more perfect and lifelike.

The cutting edge technology behind Sora

To produce realistic videos , Sora must be able to understand the user's language, requests, and commands. But it doesn't stop there, Sora must also understand the rules of physics, how things move, and how light and shadow change when the character moves.

For example, the task was to create a girl walking down a Tokyo street at night. And the output Sora created was pretty good. If you weren't a sharp eye, you wouldn't be able to spot the small flaws in the video.

The video is 99% realistic with the vividness of street lights reflecting through puddles of rainwater on the ground. These are things that are not in the command but the AI understands the physical rules of the real world and draws them on its own.

According to AI news, the technology behind Sora is partly GPT-4, a large multimodal model that can process text and image inputs. GPT-4 updates OpenAI's large language model, which is trained on large amounts of online data to generate complex responses to user queries.

The rest of Sora is a diffusion model, turning static noise into a coherent video by gradually reducing it over multiple steps. Sora also uses the same technology as Dalle-3: Turning text into images, here with incredibly realistic videos.

How will AI impact the creative industry?

People working in the fields of photography, cinematography, and graphic design are very interested in the development of AI. In Hollywood, the world's film capital, experts are already measuring the impact of AI.

According to the report, 21.4% of film, television and animation workers in Hollywood are expected to be replaced by AI by 2026.

Last summer, Netflix began recruiting for an AI product manager position with a staggering salary of $300,000 to $900,000, dozens of times higher than the salary of an average, unfamous actor.

Concerns about the uncontrolled development of AI

The AI trend is like a big wave. If you learn to surf and stay on top of it, chances are you and your job and position will be safe, and even thrive.

However, with opportunity comes risk. These super-intelligent artificial intelligence programs also pose risks. And with AI’s unimaginable development, these risks are even more worrisome.

According to Dr. Tran Van Canh, a data scientist at the Faculty of Information Technology, Heidelberg University, Germany, although the current Sora model still has many shortcomings such as confusing left and right or not maintaining a continuous image throughout the duration of the video, the risk of Sora creating false information is entirely possible, making it even more difficult to determine what is true on the Internet.

"With the Sora application, it is very difficult to distinguish between real and fake news because the videos created by Sora are very realistic. OpenAI itself is finding limited solutions, however, we have to accept that there is no radical solution. In the near future, the application will allow users to provide images and text to create new videos, so the statements of influential figures in the world can be faked very easily, causing great consequences for the public" - Dr. Tran Van Canh commented.

Before Sora, OpenAI released two products, including ChatGPT, which can generate human-like text, and DALL-E technology, which creates "Deepfake" - a technique that uses AI to create fake audio, images, and videos. Many people have been victims of Deepfake, from the Pope, former US President Donald Trump, or recently, the incident of pop star Taylor Swift having her face grafted to create fake pornographic images.

"In fact, OpenAI is a pioneer in GenAI technology. They have a very excellent team of scientists . The budget they invest in AI models and human resources as well as machine learning is very large, so the products OpenAI creates have a very high level of authenticity" - Dr. Tran Van Canh said.

This shows that the development of AI brings both positive and negative aspects to the constant flow of the Internet. And with the emergence of Sora, the technology race is expected to become hotter than ever. Currently, large technology companies such as Meta, Google and Runway AI are also researching and developing AI technology to create videos from text and have announced a number of similar models.

According to VTV

Source