laitimes

OpenAI video generation swipes, how much impact does it have on reality? | Beijing News column

OpenAI video generation swipes, how much impact does it have on reality? | Beijing News column

OpenAI video generation swipes, how much impact does it have on reality? | Beijing News column

▲The AI video clip released by OpenAI is 59 seconds to the end, and the picture is smooth and stable. Picture/Screenshot of network video

According to the Beijing News Shell Finance report, in the early morning of February 16, Beijing time, OpenAI released the first Wensheng video model Sora, which immediately sparked heated discussions.

Sora can generate a minute-long HD video based on the prompt words entered by the user. Although it has not been used publicly, in the demo video officially shared by OpenAI, Sora can directly output a picture with multiple characters, multiple scenes, and camera movements.

Sora's powerful impact mainly comes from two aspects. First, Sora's video is too realistic, "realistic, it doesn't exist anymore"; second, Sora's video level is very high, and those who engage in film and television will be unemployed, and short videos will "completely wipe out the industry's jobs".

According to the information released by OpenAI, Sora's main functions include: generating videos based on text prompts, generating videos from static images, and filling and expanding videos.

Obviously, none of these features are new to Sora, and many other tools can be implemented to varying degrees. However, the generation is not satisfactory and is prone to errors. This time, Sora takes the quality of the generated videos a big step further.

Sora can generate videos up to 1 minute long, which is much better than previous similar tools. Not only that, but Sora videos can also be shot to the end, with multi-angle camera switching, and the object is always the same. Sora videos can also use lens language such as scenery, expressions, and colors to express emotional colors such as loneliness, prosperity, and cuteness.

In short, Sora's AI video effect is almost on par with human shooting. Therefore, although it is not an original innovation from 0 to 1, it is shocking enough.

For example, in the current era of short videos, since Sora is comparable to humans, the specialized team that produces short videos may face the danger of being laid off. You must know that for short video photography, directing, editing and other tasks, Sora can do it all by "one person", and you can't do it without swiping the screen.

In the future, Sora short videos for different purposes may soon flood various video social platforms, and their impact and changes on reality cannot be described as huge.

OpenAI also announced in its announcement that Sora will be a major milestone in the realization of artificial general intelligence (AGI).

But no matter whether Sora "understands" the world or not, and no matter how realistic Sora's video is, it's still a digital analogue, not a reality shot of reality.

As the saying goes, seeing is believing, and hearing is false. However, in the world constructed by AI short videos, it is obvious that what the eyes see cannot be used as a basis for judgment.

The application of SORA will undoubtedly further exacerbate the post-truth situation in contemporary society, and the boundary between the real and the virtual will be further blurred, or even completely dissolved. This requires a high degree of vigilance.

OpenAI video generation swipes, how much impact does it have on reality? | Beijing News column

▲ In the AI video released by OpenAI, the close-up of the character's face is also very detailed and realistic. Picture/Screenshot of network video

To put it simply, the use of various information and communication technologies and artificial intelligence tools has made people in an environment where it is difficult to distinguish between truth and falsehood, and over time they have increasingly ignored the so-called objective facts, and have increasingly resorted to personal emotions, emotions, beliefs, and preconceptions in thinking, handling, and evaluating problems.

On the Internet, the post-truth situation is especially evident. Hired cyber trolls are raging on the Internet, and all kinds of rumors and conspiracy theories are emerging one after another. It's more of a baseless, casual remark, full of emotional standing, scolding, and scolding, completely impatient and unwilling to figure out what the truth is, "just do it". As a result, the network news is constantly reversed, and the plot is outrageous. The emergence of Sora could further exacerbate this situation.

Of course, pointing out the social risks that Sora could lead to is not about killing it with a stick. For the development of new technologies, society should be inclusive enough. However, before it can be rolled out to society on a large scale, it is necessary to fully anticipate the risks and challenges and be prepared to deal with them in advance.

As far as the post-truth issue is concerned, first of all, everyone must be aware that we are living in an era where reality and fiction are intertwined. Receive all the information you receive, develop the instinct of questioning, reflection and criticism, and don't listen to the wind or the rain. In addition, if necessary, the information sources should be carefully verified, compared, and analyzed to improve the ability to distinguish between the true and false information.

For relevant enterprises and industries, it is necessary to continuously advocate industry self-discipline and self-examination and self-correction, and fulfill the review obligations of content providers. The government should refine the regulatory measures in accordance with the relevant regulatory requirements, such as prominently labeling content as AIGC (Generative Artificial Intelligence), or classifying it as entertainment or advertising, without confusing it with facts and news.

As for the emergence of Sora, it may lead to the unemployment of a large number of employees in film and television and other related industries, and it is a major social risk that cannot be ignored.

In fact, there has been an in-depth study on the unemployment problem that may be caused by the large-scale application of AI, and many solutions have been proposed, including delaying the application of AI technology, helping reemployment, and levying AI taxes. Now, Sora's overnight swipe reminds us that we need to implement these measures steadily and refined, and we can no longer just stop at "swiping".

Written by Liu Yongmou (Wu Yuzhang Chair Professor, Chinese Renmin University, Researcher, National Academy of Development and Strategy)

Editor / Chi Daohua

Proofreading / Liu Yue

Read on