V.I.D.E.O.L.O.V.E - Reach more people with synthetic videos

Users use video content to find information quickly and easily. Around 86% of online content is video and videos are shared more frequently on social media. It is predicted that for 2022, the average daily video consumption will be around 100 minutes. So, publishers and businesses can't ignore video content these days. However, it is laborious to prepare written content for print and online, and as video as well. Outsourcing video production is an alternative, but it also involves high costs. Without video, however, communications companies run the risk of losing their target audience.

Attention and better rankings

There is a lot more work behind a two-minute video than is obvious at first glance. In addition to the hunt for a suitable location and extras, the shooting, setting, editing and post-production take up a lot of time. In the end, a two-minute video clip may take several workdays. Anyone who counts on video in corporate communications needs a good team, a lot of time and a sufficiently large budget. Especially for startups or small companies, the human and/or financial resources quickly exceed the limit. However, not using video is not an option, because it has an extreme impact on the company's image and is getting more and more important. Videos attract attention, can deliver information easily and have a positive effect on search engine rankings. So, anyone who can't afford video production is in a quandary.

Video creation via text

The two developers Julia Leduc and Kay Law have recognized the problem and are currently working on a solution in the Media Startup Fellowship: "We want to make video creation easy and feasible for everyone," says Julia, co-founder of Anymate Me. Together with Kay, she has developed an AI model that generates synthetic, narrator-guided videos. These narrators are animated by AI based on the user's text input. Avatars and voices are customizable. These two sample videos show what is possible, among other things.

AI videos like these enable hyper-personalization and targeted approaches to customers. Based on any text, the AI creates matching lip sync. "If any content changes, then it's only a matter of uploading the new text - the AI updates the video automatically. Obviously, that's convenient. With classic video material, as soon as something changes in the text, a new video had to be shot," Kay explains. What Kay has in mind are not only content adjustments to what is said, but also translating it to other languages. After all, what if you want to use a video in multiple languages? "Reshoot everything - not necessary. The same video can be dubbed into English and German, for example, without having to touch the visuals. The AI takes care of a matching lipsync."

Carte blanche for deep fakes?

The founders are aware that there are also dangers associated with such synthetic videos. Deep fakes, i.e. manipulated video recordings, pose the risk of spreading fake news, for example. That's why Anymate Me only offer their AI in the B2B sector, and an intelligent blocking function prevents the use of well-known faces. "Basically, we work with avatars provided by us for now. Of course, personalized avatars are also possible, but this requires the explicit consent of the person in question," says Julia.

Capitalize on the video trend

Automated synthetic video generation is an innovative approach to creating explainer videos or producing videos for social media. This makes it possible to harness the positive effects without large production costs. "The trend towards video is quite clear. In the future, we would also like to integrate more emotions and body movements like in an action scene into our AI model. The creation and synchronization of feature films is also on our agenda," Kay explains.