Generative AI is garnering enormous attention. The pace of innovation is mind-numbing.
One area that appears to be lagging is video. This is further behind its text and image counterparts. Of course, that is logical since it is a more in depth and brings on greater challenges than the other forms of media.
That said, we are now dealing with models that reason. The amount of compute being thrown at the development of these technologies is monumental. For this reason, it is practical to look at what is happening with AI video.
Let us begin by looking at the advancement and how things improved.
Here is the famous Will Smith eating spaghetti video from a couple years ago.
This is a recent compilation for Veo 2.
Obviously there is a start difference. AI video made huge strides yet, as we can see, it still is not ideal.
In this article we will dig into where this might be going.
When we cite "AI video" what do we mean?
For the sake of this discussion, we will skip the idea of videos being generated from a couple prompts. In other words, the focus is on AI video as a tool rather than the majority of the creative process.
Using this perspective, how long until we can have videos where people cannot tell the difference between human or machine generation?
This is a sliding scale question because it depends upon the skillset of the individual. This tool, in the hands of an expert will likely net better results than a novice. That is true for in any situation. The key is AI helps to close the gap.
It doesn't take an expert in video design to notice a huge difference between films made in the 1980s or 1990s compared with today. What is rolled out now is more advanced. This is due to the tool video creators have at their disposal. Generative AI will advance this a great deal further.
The contrast between the two videos above is a two year window. What is being questioned is what will happen in the next 2 years?
When looking at this, we have to decide whether the pace will accelerate, remain the same, or decline. Naturally, this is impossible to know since none of us can see into the future. My basic assumption is that, even if we see a slight decrease in the rate of advancement, this will still make the leap from the second video look measurable to the first.
At the core, Web 3.0 is a further fragmenting of different industries. When it comes to finance, we see DeFi offering options outside the traditional system. Even though it is nowhere close in size, we can see how each DeFi transaction is one away from Wall Street.
The same is true for content creation. Anything that is produced outside the scope of Web 2.0 is simply an alternative. Each article read or video viewed is a shift.
Naturally, AI video generation is not exclusive to Web 3.0. In fact, it is not just an online tool. We are already seeing major Hollywood studios incorporating them into their works.
That said, Web 3.0 should benefit simply due to the sheer multitude of content that can be created. Here is where the fragmentation enters. YouTube is going to see an explosion yet that is not the only game in town. It is seeing competition from other Web 2.0 platforms. With costs coming down, it is likely all platforms will have video as a part of their offering.
Here is where the type of video is crucial. There are different levels of quality which starts to drive the point home.
When we mention video, the natural tendency is to think of television or movies. This is the most traditional videos creation, dating back more than 100 years. For much of that time, it was the only game in town.
This changed with the introduction of the Internet. That broke up the monopoly on information distribution. Of course, text was the first disruption since it was the easiest media to deal with. Video suffered a delay due to the lack of bandwidth, high storage costs, and lack of tools.
Over the last 15 years, we saw a change. No longer is this the case. We have video all over the Internet. YouTube is the top streaming platform. Other social media sites are filled with video. Advertisers focus as much on these sites as they do on broadcast television.
When talking about films or television programs, quality matters. We are looking at the need for cinematic level productions. People will be disappointed if that is not the case.
Within this realm, there is some range. When looking at something like a Western, that is radically different than animation. The latter is more aligned with AI since it is completed generated already. It should be so surprise this is one of the first areas to see penetration by some of the AI products.
Even if cinema quality video is years away, what about the rest of the video world? Again, when we think about what we see on a daily basis, the scope is large.
Consider online video advertisements.. What type of quality do we expect with them? Do they have to rival that of Hollywood films? Of course not.
My forecast is most of the advertising videos that we see only will be AI generated by the end of 2025. This is due to a combination of the expected advancement, the comfortability of creators with these tools, and people's expectation levels.
The same is true for most YouTube videos. What is the expectation for a 10-15 video? Since the user is not investing 2 hours (nor money), the same quality level is not required. This is where the idea of "social video" comes in. We simply are conditioned to get a lower quality product when not investing much (time or money).
I am going to forecast that 2026 will be the year we see AI video take over. Please bear in mind this does not mean that Hollywood films will be exclusively made using these tools. For that realm, it will still supplement what is being done.
Instead, my view is that most online video content will be done using AI video generation. We will see another "Turing test" passed, as most people will not be able to tell the difference. At this level, they will likely not care.
People sitting in front of the camera talking about different topics or conducting interviews will not change much. There can be an enhancement of the videos using some tools but will not be generated. Other types of videos, however, could see an increase (such as educational videos).
All of this will lead to a massive shift. Just like we are seeing chatbots take away from search engines, AI video is going to disrupt things.
