I developed a digital avatar of myself that you can interact with.
Image Credits:Synthesia / TechCrunch
The Future of PR: My Experience with AI Avatars
This summer, I was taken aback when Alexandru Voica, the head of corporate affairs at the video-generation startup Synthesia, shared a link showcasing their latest addition to the PR team: an interactive virtual avatar of himself. This avatar was designed to answer common inquiries about Synthesia, such as its functionality and offerings. This interaction came just a day after a panel discussion I participated in where PR professionals inquired about my feelings regarding AI-generated pitches. However, Alexandru’s avatar represented a leap forward—an impressive culmination of AI’s potential in the field of public relations.
Synthesia: Pioneering Digital Avatars
In September, I had the opportunity to visit Synthesia’s new office in New York. The U.K.-based startup has gained notoriety in the digital avatar sector, joining innovative companies like D-ID, HeyGen, and Colossyan. Remarkably, Synthesia reached a $4 billion valuation earlier this year and announced it surpassed $100 million in annual recurring revenue (ARR) last year.
Synthesia specializes in creating interactive training videos with AI avatars, recently launching a groundbreaking product called Roleplay Sessions. This tool enables employees to rehearse sales pitches with an AI avatar that provides real-time feedback and scores their performances.
During my visit to the office launch, I was invited to create my very own AI avatar. Without hesitation, I accepted the offer. After all, I felt confident that my outfit and hair were in great shape that day!
Discovering My Digital Twin
Before encountering my digital twin, I held a somewhat neutral view of avatars, feeling they would eventually become commonplace in online platforms. With many individuals on Instagram creating avatars in their likeness for social media content, I recognized the trend as fascinating. Therefore, I was intrigued to present my digital twin, a first for Synthesia, which had not previously created avatars for journalists or anyone other than Voica. Trained on my article about venture-backed startups and their propensity for fraud, my avatar would only answer questions regarding that specific content.
To interact with my avatar, simply click “start in a new window.” You’ll be able to ask inquiries such as:
- Why did I choose to write this story?
- What insights did the research paper provide?
- What were the researchers’ conclusions?
Crafting My Avatar
To create my avatar, I entered a mini film studio within Synthesia’s office, where the team took numerous photographs and recorded two minutes of my voice. After consenting to the avatar’s creation, my “digital Dom” was brought to life. Synthesia developed a personal avatar for me that merely reads any script I provide, with options for both glasses-on and glasses-off versions. Additionally, they crafted two interactive avatars capable of conversing and responding, also available with and without glasses.
We trained my interactive avatar on a selected article, harnessing a blend of voice-to-text, video, language, and text-to-voice models. The tech stack employed for my avatar includes Synthesia’s proprietary video and voice models, though clients can opt for alternatives from other suppliers like Cartesia, ElevenLabs, Google, or OpenAI. Companies can host their avatars on their preferred cloud service or pay Synthesia for hosting.
The voice-to-text model converts spoken words to text, while an agentic language model interprets the text and takes actions accordingly. The text-to-voice model transforms responses into audio, and finally, a video model animates the avatar as it speaks.
Product Offerings by Synthesia
Synthesia creates three main types of products:
- A video-creation and distribution platform with classic avatars, where users input a script for the avatar to recite.
- An agentic platform called Sessions, facilitating interactive experiences through surveys and role-playing with avatars.
- An API platform that allows customers to integrate Synthesia’s video and voice models with other tech services to develop interactive avatars and other products.
The avatar creation process took the Synthesia team a few days. Initially, I tested the personal avatars by typing a simple script about the changing seasons in New York, my favorite time of year. The AI voice delivered a remarkably accurate rendition of my words, thankfully avoiding the hoarse quality from my original recording.
When I shared the personal avatar with some less tech-savvy friends, they found it both captivating and slightly unsettling.
The Interactive Experience
Next, I revealed my interactive avatar, which operates on a deterministic model—meaning it can only respond based on its training. In this instance, it was trained solely on my article about venture fraud. As I posed questions regarding my prior employment and my New York residence, I consistently received redirections back to the story.
While my friends noted the voice did not match mine perfectly and felt the likeness was weaker than the personal avatar, they acknowledged its eerie realism. My mother’s reaction was enthusiastic, labeling it “amazing.” She and my father playfully attempted to quiz the avatar on details only they knew about me, but it always redirected them back to my venture fraud story.
“I don’t remember giving birth to two of you,” she humorously commented after testing the model.
Reflecting on the Future of Journalism
This experience prompted me to ponder the future landscape of journalism. Would audiences accept news delivered by an avatar? One investor I spoke with firmly dismissed the idea. With the current apprehension surrounding AI-generated content infiltrating media channels, many remain skeptical. Others raised valid questions: Could avatars enhance or even replace journalists? Would corporate leaders prefer to communicate with an AI version of a journalist over a human counterpart?
For me, the essence of journalism lies in connecting with people, crafting narratives, and exploring new subjects. Yet, trust remains the foundation of this profession—something that doesn’t appear easily transferable to AI.
The Appeal of Digital Clones
Outside the realm of journalism, I acknowledge the allure of self-cloning. Imagine not having to catch up on work post-vacation because a version of you is perpetually present, fielding inquiries.
As corporate America begins to adopt avatars, my feelings remain mixed. After the initial novelty wore off, I scrutinized my avatar closely when it finished speaking, waiting—perhaps for it to blink, offer something novel, or acknowledge my presence.
However, because my digital versions are deterministic, they won’t do any of that. Yet, I can understand how someone could develop a slight case of AI psychosis with an avatar powered by a free-thinking chatbot.
I expressed to an investor that I believe my generation, Gen Z, likely won’t acclimate fully to digital twins. They seem ripped from the pages of science fiction, reminiscent of scenes from movies we have watched. Yet, I find digital avatars less jarring than humanoid robots. With avatars, if things become strange, I always retain the option to log off.
In the meantime, I present my personal avatar, delivering the latest top stories on our website this week!
Thanks for reading. Please let us know your thoughts and ideas in the comment section down below.
Source link
#created #interactive #digital #avatar #talk
About The Author
