Generative AI is an artificial intelligence model that, when trained on large datasets, can predict the next word or pixel to generate text, images, audio, and video. The simplest input to generative AI, called a prompt, is a text description. Based on that text description, a generative pretrained transformer (GPT) writes paragraphs, text-to-image models like Stable Diffusion create images, MusicLM creates music, and Imagen Video creates videos. can. This technology democratizes all kinds of content creation. When it comes to video creation, it can level the playing field more than smartphones and social video platforms already do. It will also fundamentally change the video content industry.
Consider the stars of this domain: Netflix, TikTok, and YouTube. While each is unique in terms of content types and business models, all three platforms motivate creators to develop compelling content, match the right content to the right consumers, and identify which content It is run by identifying what drives engagement. Each of these elements built upon each other to create a flywheel that helped all three platforms gain audiences at high speed. But that flywheel is starting to lose momentum. Generative AI exacerbates the problem by creating a new video content creation value chain.
Why Netflix, Tiktok, and YouTube are in trouble.
Netflix, TikTok, and YouTube do well due to their ability to determine content relevance and engagement. They all have a huge amount of data about who watched what and how. Despite their success, determining the “what” still presents two serious challenges.
Useful and accurate feature extraction. When a video is commissioned (as is done with Netflix), categories such as genre, performers, and length are known. Of course, many of the features of the video can be specified. Scripts, shot lists, and other production features are pinpointed. But trying to use this data presents another extreme problem. It might be too much information to describe just one video.
Overcome barriers to creativity. Closed Hollywood-style content production is expensive and time consuming. Netflix will spend $17 billion on content in 2022. Netflix Co-CEO Greg Peters said: Wednesday every week, glass onion You can win back the majority of those viewers every week. ” Obviously they can’t deliver yet Wednesday (a popular high-budget modern spin-on the adams family) every week using the current production model.
An alternative model is the creation of open user-generated content used by TikTok and YouTube. While this is relatively cheap and quick, it has three (sometimes conflicting) goals: 1) retain influential creators, 2) motivate new creators, and 3) maintain and grow your audience base. Incentives should be set to balance Platforms in this space are waging an incentive war as they attempt to generate a sufficient amount of compelling content from a relatively small number of popular creators. It is said that YouTube Shorts, on the other hand, have lowered the bar for creators to make money. Instead of his TikTok requirement of at least 100,000 followers, he only wants 1,000 subscribers.
These two challenges partly explain the failure of short-lived streaming platform Quibi. Quibi combines all three weaknesses into one platform. It doubled down on a closed Hollywood-style production system by hiring expensive creators and actors. Instead of empowering individual creators like YouTube and TikTok, Quibi bet on brand-name creators and actors. What I got in return was poor (perhaps second tier) content that didn’t work. That’s because while targeting millennials and his Gen Z, they’re not cultivating creators for that generation. Also, to my surprise, it didn’t use AI to decide what content to create (although it did use AI to recommend what to watch for viewers).
No human-driven platform has yet overcome both of these challenges. However, a solution may exist. Generative AI will change what video content you create, how you create it, and who you see it with, bringing a whole new kind of AI-enabled platform.
Towards a generative platform.
Imagine this scenario. The author enters the following text description: Two people are sitting in an Art Deco cafe. It’s snowing outside. One of them nibbled at her Swiss wedge of cheese and said, “I’m creatively constipated!”
Hyper-realistic live-action videos (with sound) are generated almost instantly and displayed to billions of viewers. Know who watched how long, who skipped which parts, likes, shares, comments, searches, and all off-platform discussions about the video, as well as used to create that video I also know the exact input given. This scenario solves his two challenges of existing video platforms in one go. It provides a more accurate description (input text prompt) of the video and significantly lowers the barrier to creation (as easy as entering your imagination). Don’t worry about CapCut or actors.
It sounds like magic, but it doesn’t actually exist yet, it’s just a combination of three AI programs. AI #1 generates videos based on text input. AI #2 matches videos with the right audience. AI #3 uses the resulting engagement to guide creators on what to create next. A more primitive version of this production model is already producing content, perhaps most notably using generative AI to create the script and having nearly 100,000 followers parodying Seinfeld’s sitcom ” Nothing, Forever.”
Generative AI-powered video platforms reduce barriers to value creation by guiding creators on what drives engagement and showing viewers relevant content. At the same time, fewer barriers and improved guidance help creators increase the value they can create outside the company. And with near-zero friction between creating and viewing relevant content, creators are also viewers and vice versa. As the viewer types in a search, the boundaries blur further and that input text becomes the prompt for the new video.
The economic impact will be huge. Traditionally, a small portion of the very popular content on the platform made up for the majority of the less popular content. Generative AI platforms drive the success of popular content. This is because creators are enhanced with the help of algorithmic recommendations on what to create next. At the same time, much lower barriers to creation make the rest more profitable.
How will the major existing platforms adapt? Of the three, Netflix is the most constrained by its business model and probably difficult to change dramatically. I’ve resisted the model, but only recently moved in that direction. TikTok comes closest to a generative AI platform in terms of its business model, features and flexibility with regards to future plans, but it is under regulatory scrutiny in the US. YouTube is well-positioned as it tries to beat the competition by introducing short videos and improving creator incentives. It also backs Google’s AI capabilities. But Google has already shown that it is slow to move commercially in the generative AI space.
These were just the opening credits.
The recent technological advances and growing awareness of generative AI have been astounding. Indeed, there is no technology yet to generate hyper-realistic live-action video from text input. The availability of such technology is key to enabling new platforms.
Even when text input becomes available, it often fails to provide a sufficiently precise definition of a video, and as creators and viewers create similar text, platforms are forced to produce large numbers of similar, but not identical, videos. You may see it generate a . And as platforms learn the keys to producing compelling content, how will conflicts of interest between platforms and creators be managed? How do we prevent the inevitable propaganda and misinformation that spreads?
Despite these caveats, generative AI will very likely power new video content platforms that will replace, or at least complement, current Netflix, YouTube, and TikTok. Generative AI technology is used not only to create content, but also to enhance platform dynamics between platforms, creators and consumers. Of course, none of this would be possible without technical uncertainty and ethical risk. And video, of course, is just one area where we can expect such rapid change. Many other creative areas, such as art, music, and letters, are driven by dramatic changes and people who can discern what is ripe for destruction, or take advantage of generative AI to take control of their turf. We are about to welcome new business opportunities for those who protect us.
