Recently, OpenAI released the first **generative model SORA, and showed several short demos of Wensheng**, which caused a wave of discussion, almost all of which pointed to the changes brought by SORA to the media industry, and Musk also said on social **: gghumans (human bets admit defeat). How will SORA change the media industry? What are the implications for the average person? In this regard, we interviewed Shen Hao, chief scientist, professor and doctoral supervisor of the State Key Laboratory of Convergence and Communication, Communication University of China.
World Simulator".
At present, the effect of SORA is far superior to that of previous AI models, how well is it capable of handling complex scenes, character movements, background details, etc.?
SORA is a new phenomenon-level application scenario of large language models after GPT. Its characteristics are mainly visual, especially its ability to output a surreal state for one minute in one shot, so it brings a great intuitive feeling and impact to everyone. After the emergence of this product, it has aroused a great perception of innovation in the whole society, whether it is the industry or the whole public. I think first of all, we are happy to see that AI is constantly innovating and developing.
AI** clip released by OpenAI Source: Network.
By now, we have seen that SORA may have produced hundreds of works, and its innovation and ability are reflected in many ways. The first is its one-shot camera movement and perception of the lens, and the second is the content it generates, which some people call "world simulator". In fact, this concept is forward-looking and futuristic, that is to say, it is not only in line with our human perception of the physical world, but also can reconstruct our perception of the physical world. So the work it generates has a very strong visual impact, and it's not even something that we can achieve by shooting. Finally, the content generated by SORA provides unlimited imagination for everyone.
According to industry insiders, SORA will shorten the implementation of AGI from ten years to one or two years, what do you think?
The full name of AGI is General Artificial Intelligence, and SORA belongs to AIGC, that is, the use of artificial intelligence to generate content. AGI not only refers to the digital world, in other words, not only visual or film and television works, but also robots, unmanned driving, and even broader application prospects. From this point of view, SORA is not the core part of AGI development.
However, because of the emergence of SORA, AGI may have more tangible perception, such as artificial intelligence robots, physical robots or humanoid robots, and of course, intelligent driving cars in the future. Therefore, there is no equivalence between SORA and AGI, but it will help AGI to understand the world and understand reality.
"People" in artificial intelligence are the most important.
Could SORA be an important tool for the media industry? How will it affect the media industry?
First of all, SORA requires a lot of computing power, so it is still difficult to promote and popularize it.
Secondly, if we make an attempt, such as making a movie or a 60-minute show with Sora, it will take minute by minute to generate. On this basis, how to connect the story smoothly is a big problem. If you look closely at the ** generated by SORA, there are still some flaws in the entire space, just like AI painting, which cannot fully reproduce the reality. Therefore, as far as the media industry is concerned, the content it generates mainly has a great impact on the fields of self-development, short-term education, entertainment, education or science fiction.
Source: Qianku.com.
Many people think that SORA will affect their future career development, but in fact, it is normal for the emergence of an innovative technology to trigger some imagination of the future. Not only SORA, but since the birth of AIGC, it has been influencing people's imagination of future careers. The impact must be there, it will weed out some people in the industry, but it will not weed out good journalists or creators. Because after all, we still have to use text to generate creative ideas, which is what humans need to do, but it makes our ideas concrete. For the film and television industry or the media industry, it is another means to increase production and efficiency.
What does the future hold?
What are the current weaknesses of SORA? What are the challenges in practice?
First of all, I prefer to say that the ChatGPT-based large language model itself is a subversive and revolutionary product, and SORA is just a new node in the continuous iterative evolution of this product, which is the evolution and extension of ChatGPT's technology.
Source: Qianku.com.
In addition to the cost of the application and the computing power generated, I think the disadvantage of SORA is the problem of technical misuse. The so-called misuse is not only false information and falsified information in the news, but also criminal fields such as telecommunications fraud and pornography that have been applied since the emergence of artificial intelligence.
If these digital works have a greater impact on the social value system in the future, or even illustrate a possible scenario, it may make people have a wrong social perception. So now all the big models are severely restricting these issues, such as **, hate, race, and of course, the cognitive denigration of certain groups, and I believe that the big models will achieve value alignment on this issue.
From ChatGPT and Midjourney to the current SORA, what impact will these AI technologies have on the lives of ordinary people?
In fact, the more automated and intelligent the technology, the more civilian it will be. What used to require professionals to complete, now many students and even middle-aged and elderly people will play, for example, now people's mobile phones ** a translation app, do not need to master his Chinese language can go abroad, to achieve the switch between hundreds of national languages. Similarly, if you do self-certification, it used to be very difficult for you as a non-professional to make a movie-level or more professional-level product, but if SORA is open-sourced in the future, or other companies release products like SORA, then it will become very easy, including now that there is a lot of such content in many short ** platforms.
So my point is to embrace technology. It's not difficult, it's becoming more and more instrumental, it helps us to be more productive, it compensates you creatively if you're lacking in ideas, and it can at least give you hints if you're struggling with content production. But if you don't embrace it, you might really be eliminated.