With the launch of Sora, AI in video creation has finally reached a level where even the skeptics have started to pay attention. You might remember the first AI generated videos, including this jewel.
Well. Times have changed. The newest videos that OpenAI shared this week look more like this:
Needless to say. It’s time for content creators and marketers to pay serious attention. Sora is currently still in beta and can only be used by a small number of early access testers. However, knowing OpenAI, we can safely assume we will all get our hands on this sooner rather than later.
Regardless, there are already many AI tools available for marketers / content creators when it comes to UGC. Let's explore the most common AI use cases:
Most Video editing software today offers creator simple video editing tools powered by AI.
Prominent examples are eye contact correction, stabilization of shaky videos, video background removers and automatic reshaping and resizing tools.
One limitation we have found is that we are still missing a tool that allows creators to upload lots of B-Roll and receive a final cut video at the end that matches the voice over. Meaning, our creators still cut their videos themselves.
Ideally, we would even see a variety of cuts from one set of B-Roll which has become quite common for still ads.
AI powered transcriptions are almost flawless today (except for Swiss German), allowing subtitles in all languages with ease.
On the voice side we can generate all voices in all languages from text to speech but these are still recognizable as “AI”. There are tools (like voice dubbing from veed) that are trying to address that. I have tested it for you here (and was actually quite impressed).
Another neat tool regarding voice is the Loom edit by transcript. Where you can delete items in your voice-over by simply deleting them off your transcript.
Some UGC videos aren’t just shot footage, they also include stock videography or photography. Creators can use AI generated videos and photos for these purposes. This is also where Sora will have the most obvious application for UGC creators.
Now to the meaty one… Can creators / marketers already create UGC fully with AI today? Yes. Does it show it’s AI? Also, yes.
One option is that they just use an Avatar. These are widely available and can look something like this:
The best out there I have seen comes from my dear friend Mathis Büchi Ginzbourg with magicads.ai. An example of what they can do here:
I'm willing to make this my Steve Ballmer moment of when he laughed at the IPhone:
But I do believe authentic User-Generated content is here to stay. Not despite AI UGC, but because of it.
Do I think that AI will get massively better and create UGC videos that are hard to distinguish from the non-AI versions? Yes.
But in a world where AI does everything I believe that consumers will want authenticity and creativity above all else. And UGC will remain the key way to obtain that, as long as it manages to communicate its authenticity message loud and clear. (If it doesn't, it will cease to perform)
The way I see it, UGC is a bit like an onion. In its core, the authenticity must remain for it to perform. It's the key reason it started to perform in the first place. But the layers above, the processes, the editing, the scalability - all these should be enhanced and enriched with AI wherever possible. We will want to use AI to make our creators & our own life as easy as possible.
