5 Themes From the Synthedia Conference
The trajectory of synthetic media from 10 industry innovators
Synthedia the conference — the first ever synthetic media conference covering the entire scope of the technology — is in the books. Yesterday, over a hundred people gathered to learn about the innovations and from the innovators shaping the synthetic media market. Hundreds more signed up for access to the recordings. Those recordings will be made available on Voicebot’s YouTube channel over the next three weeks. Stay tuned!
However, why delay? We are in an instant gratification era, so Eric, Andrew, and I thought we’d offer our Synthedia conference hot takes.
Synthetic Media is Not One Thing
Andrew Herndon says, “Synthetic media is not limited to one specific media but rather is a collection of media types empowered by deep learning models and decades of human-computer interaction research.” Agreed!
We heard from Anne Spalter (Anne Spalter Studios) who is using AI-powered image generators in her art. Brandon Kaplan of Skilled Creative shared a short essay written by GPT-3 and presented by a virtual human about the Luddites of the synthetic media era. I used a different virtual human to recite a poem that I generated using a GPT-3 prompt.
Jean-Baptiste Martinoli debuted his 3-minute film about an intelligent alien race that sent us a message. Everything except for the script was developed using AI-based synthetic media tools. He has a more recent short film where the script also was generated by AI.
Greg Cross from Soul Machines, Natalie Monbiot of Hour One, and YouTuber Dom Esposito talked about digital people (including digital clones) and the productivity and engagement enhancement they can deliver. Respeecher’s Alex Serdiuk and Resemble AI’s Zohaib Ahmed discussed the rise of voice clones and synthetic speech and that was not all.
I also presented a framework for understanding the synthetic media vendor landscape (more on that tomorrow). We collectively covered a lot of ground.
Synthetic Media Use Cases Are Numerous
Many people are now familiar with virtual humans and speech synthesis. Eric Schwartz notes, “Synthetic media can appear anywhere and is already proliferating across verticals that most may not realize including industrial planning, healthcare, and software design. It’s not just about entertainment deepfakes.”
On that point, Eric interviewed Paul Cutsinger from Nvidia, who discussed industrial manufacturing planning, wireless coverage mapping, and architecture applications in simulated environments. These digital twins of the existing or planned physical spaces can identify potential issues or optimization opportunities before anything is constructed.
While the entertainment use cases are the most widely known and many people are now experiencing synthetic media customer engagement solutions, there are many more applications that are emerging.
Hyper Automation and Hyper Creation
Automation and creation were key themes hit on by several presenters. Using AI solutions makes content generation faster and added creative opportunities to the process. The poem written by GPT-3 and recited by a virtual human created by Hour One is one example.
The human involvement was simply to write a short prompt of one sentence for the poem and copy the results over to Hour One’s Reals program. About five minutes of effort was expended. Brandon Kaplan discussed how his agency, Skilled Creative, is using synthetic media to augment and streamline the creative process for client projects.
Autonomous and Interactive Synthetic Beings
At the outset, I introduced a framework for understanding the synthetic media market that included a distinction between virtual beings and synthetic speech that were scripted versus autonomous. The former is dynamic but must be actively controlled by a human in pre-production or in real time, while the latter included what Greg Cross from Soul Machines called a digital brain.
Most of the synthetic media technologies today are designed for asynchronous consumption. Autonomous digital people that can interact with you in real time are improving rapidly. Greg showed digital twin examples from golf legend Jack Nicklaus and NBA All-Star Carmelo Anthony. He expects the use of autonomous digital beings will increase even faster as metaverse spaces become more popular because there will need to be a scalable way to interact with users.
Nothing Hotter Than AI-Generated Art Right Now
While the entire synthetic media industry is moving quickly right now, the AI-enabled image generation space is white hot. A pure novelty a year ago, Anne Spalter talked about her sold-out collection of 501 AI Spaceships created using Night Cafe, Brandon Kaplan discussed using Midjourney for mood boards and client ads, and Jean-Baptiste Martinoli became a filmmaker in just a few hours with AI-generated assets.
An interesting comment from Spalter was that a year ago, no one had any idea what she was talking about when discussing the use of AI in her artwork. In her current show in New York, everyone was familiar with the technology. This suggests the awareness of synthetic media solutions is rising quickly, even if the term that represents the broader market is not widely known.
Another awareness-building event has been America’s Got Talent which we’ve written about here and here. Alex Serdiuk and the Respeecher team joined us and had the opportunity to talk about resurrecting the voice of young Elvis for the show this week. Millions of viewers saw deepfake and voice cloning software on an American variety show over the past month. Millions more saw the videos on YouTube. Synthetic media is quickly becoming common in everyday consumer lives.