On January 27, 2023, ANBOUND’s founder Kung Chan pointed out in an analysis that "the real huge impact of artificial intelligence (AI) is about to come". Chan noted that, at that time ChatGPT which is based on OpenAI’s model, had become immensely popular, allowing visionary people to see a future form of digital intelligence. AI is no longer mere empty talks but is taking big strides. Faced with the impact of AI, most industries will be redefined, the position of industry leaders will be overturned and changed, and the industry will undergo a major reshuffle, forming a clear technological trend. Future products and goods must achieve super-intelligence, and they must incorporate new AI systems beyond the original, thereby establishing a new generation of products and systems.
Coincidentally, almost one year later, on February 16, 2024, Sam Altman, CEO of OpenAI, released Sora, the first video generation model. This once again astonished the world. Sora can follow text instructions to generate high-definition videos lasting up to 1 minute. From the released video content, the image quality is quite exquisite. For example, in the video of the imagined Dragon Year Spring Festival by AI, on the crowded streets, there are a large number of unique characters, including children following the dragon dance troupe who look curious, and there are also many people taking out their phones to take pictures. In one video, a trendy young woman walks on the streets of rainy Tokyo, with the reflection of the neon lights on the damp ground, the video looks as if it is part of a blockbuster movie. In yet another video, a number of paper airplanes shuttle through dense forests like birds. What is surprising is that Sora knows what will happen when the paper airplanes collide and can realistically represent the changes in light and shadow of these paper airplanes.
Image: Still from the video generated by Sora depiciting Spring Festival’s dragon dance
Source: OpenAI
Sora can now generate complex scenes with multiple characters, including specific movements, and accurately depict details of objects and backgrounds. It can create multiple shots in a single video, and rely on a deep understanding of language to accurately interpret cue words, retaining characters and visual styles. The characters created by Sora can express rich emotions, so realistic to the point of being near flawless. Sora not only understands the requests presented in the prompts but also how these objects exist in the physical world. OpenAI does not shy away from Sora's current weaknesses, such as its difficulty in accurately simulating the physical principles of complex scenes and its potential inability to understand causality. Sora may also confuse spatial details in the prompts, such as mixing up left and right, and may have difficulty accurately describing events that occur over time. OpenAI stated that it is teaching it to understand and simulate the physical world in motion, with the goal of training models to help people solve problems that require interaction with the real world.
The videos generated by Sora are not only visually stunning and detail-oriented but also represent a significant breakthrough beyond previous AI works. With a profound understanding of language, Sora can accurately comprehend the user's instructions, grasping how these elements manifest in the real world. Sora has transcended the limitations of simulating the real world in 2D space, achieving two capabilities: understanding of the real world which surpasses perceptual abilities, and the ability to simulate the physical world. Some experts suggest that if Sora could understand and learn visual data like ChatGPT comprehends and learns textual natural language, it would possess a remarkably wide range of applications.
The abilities and potential demonstrated by Sora have quickly attracted high attention from the tech industry, business sector, and investment community. OpenAI stated that generating videos based on text prompts is just one step in its overall plan. While ChatGPT achieves the breakthrough in AI's accurate understanding and generation of natural language expressions, Sora takes it a step further by creating more complex video content based on understanding natural language. Following the logic of AI development and the trend of technological advancement in AI, it is entirely believable that Sora will generate even more refined and realistic works in the future, with a more diverse range of applications in various fields. If these applications are combined with other technologies, they will have a significant catalyzing impact on the world. This is precisely what ANBOUND has pointed out on the "truly enormous impact" of AI on the world.
Firstly, there is the impact on the development of many industries and related employment. After the launch of ChatGPT, many industries began to change their operations, many professions were affected, and a number of jobs began to be replaced by AI. Similar things will happen after the introduction of Sora. Judging from the reactions online, professionals in the film and television industry are beginning to worry that the industry will be severely impacted by AI in the future, and even their livelihoods will be taken over by AI. Just last year, Hollywood screenwriters launched an industry-wide strike to protest against the widespread use of AI tools by production companies for scriptwriting. The emergence of Sora could entirely bring the AI battlefield to Hollywood and even have a significant impact on the global film and television industry. Of course, the advancement of AI applications will also create a large number of new professions. If human learning and accumulation of knowledge and skills are replaced by AI learning, the world will undergo significant changes. OpenAI's ChatGPT can pass the bar exam without attending law school, while Sora may be able to make films without attending film school.
Secondly, the emergence of Sora will make it even more difficult to distinguish between what is real and fake information. In the past, much fake information was presented through textual fabrication. Later, with the progress of AI technology, false information expanded greatly through the combination of text with voice, images, and videos. In the future, further development of AI technology will enable malicious parties to create "completely authentic fake information" through even more sophisticated means. What one hears and sees will not necessarily be truth in such a future world.
Thirdly, the advancement of AI technology and the easy accessibility of application tools may create a plethora of new ethical and legal issues. ANBOUND believes that the acquisition, creation, and dissemination of information are the "universal powers" of the information society, breaking the monopoly power previously held only by power institutions (such as states or media companies) in these areas. However, the advent of the Internet changed everything. In the online society, the decentralization and universalization of information power greatly undermined the information monopoly of power institutions. At the same time, it has also created new legal and ethical issues. The various problems in the digital age will be exacerbated by the advancement of AI technology. Imagine, in the future society, if billions of netizens can easily produce and publish videos or films using tools similar to Sora, what will the world become?
Fourth, will the giants mastering AI technology become an all-powerful monster? The advent of the internet age did not bring about true information diversity and freedom, and there is still an information monopoly and an "effective information desert" in the era of massive information. The controllers of the underlying core technologies and rules of the internet still have control over information audiences. Similar situations may also arise in the AI era in the future. For example, high-level AI large model technology and applications may be controlled by a very small number of companies, making these AI giants another form of power that is difficult to be constrained by law. If this trajectory continues, AI giants could very well become all-powerful monsters
The emergence of Sora is a strong catalyst in this AI era. Its prospects and potential impacts will have lasting effects on industries, governments, consumers, and investors. At present, we still find it difficult to provide a detailed description and analysis of the impact Sora will have on the world, but it will certainly create new crucial points of interest. ANBOUND’s founder Kung Chan has analyzed the impact of AI on the world, and now the same analysis applies to Sora. Chan believes that there will be two important concerns in the future world: firstly, the complete blending of the real and virtual worlds, where distinguishing between them will be highly valuable and necessary; secondly, the creation of content will become a hotbed of novelty, as determining what kind of content is fresh and exciting will be critical.
Final analysis conclusion:
No matter how AI develops, it cannot completely replace human thinking and creativity. Therefore, the unique territory that humans will retain when coexisting with the AI world in the future will revolve around how humans control AI and engage in creative work.
