Producing media is no longer a physical, but a digital process, meaning we can create content in completely new ways. So, as computers get smarter, there’s less need for human input in content creation, and a new category of media is emerging. We call it synthetic media.

What Is Synthetic Media?

Throughout history, media has constantly evolved—from cave paintings to print to mass media to synthetic media.

But what actually is synthetic media? How is it different from the media formats we all know? What does it mean for the future of media creation?

Let’s start with a simple definition: In recent years, the term synthetic media has emerged in common parlance as a catch-all to describe video, image, text, or voice content that computers have fully or partially generated.

Essentially, synthetic media enables people to create pictures based on text prompts and realistic talking-head videos by typing in text that tells the AI what we want.

How Does Synthetic Media Work?

So, what’s going on behind the scenes of generating synthetic media?

Artificial intelligence (AI) algorithms create synthetic media. Machine learning models recognize and analyze patterns of input data and learn to perform tasks based on these data without human intervention.

Here’s the key thing: Advances in deep learning are enabling, essentially, magic. Text, image, voice, and video content can be reproduced at such high fidelity that it’s almost impossible to tell whether it’s real or not.

But as humans, we’re hardwired to spot even the slightest errors and slip-ups in reproduced human likenesses. This is known as the uncanny valley, and it’s the reason why most digitally created characters in Hollywood movies today still feel kind of weird and why you can usually detect deepfakes.

But synthetic media technologies are evolving as you read this. It’s only a matter of time before making a Hollywood-grade film using AI will be possible.

Let’s consider video production, for example. Creating a five-minute video in the “traditional” way requires many resources. It’s often a fragmented and lengthy process that involves writing a script, hiring a film crew and actors, gathering cameras and microphones, filming the script, and editing the film into a final video. Next, think about translating the footage into different languages or updating it with new information. The average cost of a one-minute video done traditionally ranges from $1,500 to $10,000, while an average AI video generator charges $30 for 10 minutes of video content.

AI enables you to generate synthetic videos directly from your browser by simply typing your text in any language and selecting an avatar (based on a real person). You can also have your own avatar created and your own voice synthesized. Unlike traditional video, AI video is generated in minutes, and you can update it anytime.

Pretty cool, don’t you think? And we’re only talking about video here—let’s not forget text, image, and voice.

The State of Synthetic Media Today

Today, technology is constantly improving in ways that lead to new communication methods. In other words, there are frequent new media formats that vary in terms of creation, consumption, and contextualization.

Although synthetic media is still in its infancy, many companies are starting to use these solutions to create videos for various purposes, such as employee training, marketing, and internal communications. And we’ve only recently reached a point where synthetic media technologies are—put simply—smart enough to read initial sets of data and create new content based on that input_._

Bottom line: The capability of AI software to generate novel content is potentially one of the most exciting developments enabled by recent progress in deep learning. The forthcoming paradigm shift in media creation and consumption will likely alter the equation for entire industries, including education, entertainment, medicine, and retail.

For more on synthetic media, join me and my co-presenter Elly Henriksen at TK 2024 for the Creating Quality Video Content With AI session.