Types of Synthetic Media
The difference between synthetic and non-synthetic media
In differentiating between what would be considered synthetic vs. non-synthetic media, synthetic media would be any aspect of media that involved partial or entire production without human input. Non-synthetic media is considered "conventional" where a human writes an article for a blog or edits an instructional video for an upcoming training session.
Synthetic media can come through a variety of mediums:
Text-based synthetic media
The creation of written content using a combination of machine learning software and AI algorithms to generate human-like text based on user prompts (Bhat, 2023).
An example of this can be an individual using a chatbot to generate news reports.
Audio-based synthetic media
The conversion of written text into an AI generated audio clip. Synthetic audio can learn the nuances of human voices by being trained on recorded human speech for realistic responses, narration, or full length conversations.
Image-based synthetic media
The ability to generate visual content from a text-based input. This can include images or artwork to develop realistic visual representations within a few seconds.
An example is shown here in Canva for how to generate an AI image from a text prompt.
Video-based synthetic media
The text-to-video technology that enables users to develop avatars, digital characters, and AI generated images within the video content. The versatility of this media allows for users to create videos resembling real people.
This has led to the conversation about deepfakes, briefly shown here as a depiction of Morgan Freeman and explored further in Opportunties & Challenges.
Examples of synthetic media platforms and applications
An AI-based platform that combines all of the types of media described above is Synthesia. This AI video communications company promotes themselves on professional social media websites such as LinkedIn for Learning & Development tools for corporate training.
Another professional AI-based platform that allows for customization and personalization of various components of making a video while emphasizing the efficiency of the platform to make videos quickly. This includes over 100 avatars to choose from and over 40 languages to deliver products such as eLearning videos.
Developed by OpenAI, the same company that developed the chatbot ChatGPT, DALL-E 2 is a generative AI systems that creates realistic images, art, and short videos from text based prompts, presenting a versatile tool for educators in how to apply a platform like this in learning experiences.
OpenAI also announced that a DALL-E 3 is coming in late 2024, promising better text recognition for key words and greater detail in the images it creates (OpenAI, 2024)