The landscape of public relations and journalism is undergoing a profound transformation, highlighted by a recent development where a journalist encountered an interactive virtual avatar of Alexandru Voica, head of corporate affairs at the video-generation startup Synthesia. This encounter, initially surprising, escalated when the journalist subsequently consented to the creation of her own AI digital twin by Synthesia, a move that has ignited discussions about the evolving role of artificial intelligence within media, corporate strategy, and the very concept of professional identity. This event transcends the debate around AI-generated text in pitches, introducing a complex new dimension to human-AI interaction in professional spheres.
Synthesia’s Ascent in the AI Avatar Market
Synthesia, a prominent player in the burgeoning field of digital avatar technology, has rapidly established itself as a significant force. Originally founded in the U.K., the company has expanded its operations, including a notable new office space in New York. Its trajectory reflects the accelerating investment and innovation within the AI sector, particularly in generative AI applications. Earlier this year, Synthesia achieved a remarkable $4 billion valuation, underscoring investor confidence in its technology and market potential. Furthermore, the company reported crossing $100 million in Annual Recurring Revenue (ARR) last year, signaling robust commercial success and adoption of its enterprise-focused solutions.
Synthesia operates within a competitive yet rapidly expanding ecosystem of digital avatar startups, alongside other notable entities such as D-ID, HeyGen, and Colossyan. These companies are at the forefront of developing sophisticated AI-powered platforms that enable enterprises to create realistic and interactive avatars for a multitude of purposes. Synthesia’s core offering empowers businesses to build interactive training videos utilizing AI avatars, streamlining educational processes and enhancing employee engagement. A recent innovation, "Roleplay Sessions," further extends this capability by allowing employees to practice scenarios, such as sales pitches, with an interactive AI avatar that provides real-time responses and scores, offering a dynamic and personalized learning experience. This innovation moves beyond passive video consumption, ushering in an era of active, AI-driven coaching.
The Genesis of a Digital Twin: A Journalist’s Experience
The journalist’s journey into the world of digital twinning began during the opening of Synthesia’s new New York office. Presented with the opportunity to create her own AI avatar, the decision was made without hesitation, driven by a blend of curiosity and an acknowledgment of the inevitable integration of avatars into daily online life. The initial appeal was also personal, with the journalist noting a particularly favorable outfit and hairstyle on the day of the recording, a minor yet human detail in an otherwise technologically driven process.
The creation of this specific digital twin marks a significant milestone: it is reportedly the first time Synthesia has developed a digital avatar for a journalist, or for any individual outside of its corporate affairs head, Alexandru Voica. This particular interactive avatar was trained on the journalist’s published story about why venture-backed startups commit more fraud than non-VC-backed startups. Consequently, its responses are strictly confined to this specific subject matter, making it a deterministic model.
The technical process for generating the avatars involved a visit to a specialized mini film studio within Synthesia’s office. Here, numerous photographs were taken, and a two-minute voice recording was captured to build the foundation of the digital likeness and voice. Explicit consent was obtained for the creation and use of these avatars. The outcome was a suite of digital representations: a personal avatar capable of reading any script provided (with and without glasses), and two interactive avatars (also with and without glasses) designed to engage in conversation based on their training data.
The interactive avatar’s technical architecture is a sophisticated blend of various AI models. It is powered by a combination of voice-to-text, video, language, and text-to-voice models. Synthesia leverages its proprietary video and voice models, while also offering customers the flexibility to integrate alternatives from other leading AI labs such as Cartesia, ElevenLabs, Google, or OpenAI for specific components. Enterprises also retain the choice of hosting their avatars on their preferred cloud infrastructure or opting for Synthesia’s hosting services, providing flexibility and control over their digital assets. The operational flow involves a voice-to-text model transcribing user queries, an agentic language model interpreting the text and formulating a response, a text-to-voice model converting this response into audio, and finally, Synthesia’s proprietary video model animating the avatar as it speaks, creating a seamless and lifelike interaction.
Exploring the Capabilities: Personal and Interactive Avatars
Synthesia’s product portfolio is broadly categorized into three main types. First, a video-creation and distribution platform featuring "classic" avatars, where users input a script, and the avatar delivers it. Second, an "agentic platform" known as Sessions, which facilitates interactive engagement with avatars for purposes like surveys or role-playing scenarios. Third, an API platform that allows developers to integrate Synthesia’s video and voice models with other technological services to build custom interactive avatars or other AI-driven products. This modular approach highlights Synthesia’s ambition to be a foundational layer for AI-powered digital human interactions.
The creation of the journalist’s avatars took the Synthesia team a few days. Initial experimentation focused on the personal avatars. A generic script about the arrival of fall in New York was used to test the AI voice. The voice fidelity was noted to be remarkably accurate, successfully replicating the journalist’s vocal patterns while conveniently omitting any hoarseness present during the original audio recording. This immediate impression of realism, coupled with subtle improvements, underscored the technological prowess involved.
Reactions from non-tech-savvy friends were a mix of intrigue and apprehension, describing the experience as "interesting and creepy." The interactive avatar, however, presented a more constrained experience due to its deterministic nature. When asked questions outside its training data – such as the journalist’s prior employment or residential area – the avatar consistently redirected the conversation back to the venture fraud story it was programmed to discuss. While some friends felt the interactive avatar’s voice and likeness were less precise than the personal avatar, it was nevertheless close enough to evoke a "somewhat creepy" sensation. The journalist’s mother, however, offered high praise, deeming it "amazing" and attempting to ask questions only a parent would know, only to be met with the same redirection to the programmed topic. Her humorous remark, "I don’t remember giving birth to two of you," encapsulates the uncanny yet intriguing nature of digital twins.
AI’s Expanding Role in PR and Corporate Affairs
The introduction of advanced AI avatars into corporate communications, as exemplified by Synthesia’s own AI press officer and the journalist’s digital twin, marks a pivotal moment. For public relations, the implications are vast. While the initial debate centered on AI-generated text in pitches, the shift to interactive, visually realistic avatars represents a significant leap. These avatars could potentially revolutionize how companies disseminate information, manage media inquiries, and engage with stakeholders. Imagine a global corporation having AI avatars of its spokespeople available 24/7, capable of answering frequently asked questions in multiple languages, ensuring consistent messaging and rapid response times. This could augment human PR teams, allowing them to focus on more strategic and nuanced communication challenges.
However, this advancement also introduces new complexities. The ability of an AI avatar to answer questions about a specific topic, like Synthesia’s AI press officer, streamlines information delivery but also raises questions about the depth of interaction and the capacity for spontaneous, unscripted dialogue that human interaction often requires. The journalist’s avatar, limited to a single story, highlights the current deterministic nature of some of these advanced models. While beneficial for controlled information release, it contrasts sharply with the multifaceted nature of human journalistic inquiry.
The Future of Journalism: Trust, Augmentation, or Replacement?
The experience of creating and interacting with a digital twin has inevitably prompted a critical introspection into the future of journalism. The question of whether the public would accept news presented by an AI avatar is contentious. An investor immediately dismissed the idea, reflecting a broader sentiment of skepticism towards AI in sensitive sectors like news. This skepticism is further fueled by ongoing public pushback against "AI slop" – low-quality, often misleading, AI-generated content that has infiltrated social media and other news-sharing platforms, eroding trust and credibility. Instagram and Pinterest, for instance, have begun implementing controls to limit undisclosed AI profiles and content, acknowledging the growing concern.
Yet, opinions within the industry are divided. Some wonder if avatars could augment or even partially replace human journalists, especially for routine reporting or information dissemination. The scenario of CEOs preferring to interact with an AI avatar of a journalist over a human one, potentially to control the narrative or avoid uncomfortable spontaneity, presents a chilling prospect for the principles of independent reporting.
At its core, journalism is built on connection, storytelling, research, and, most importantly, trust. The journalist unequivocally states that trust, a cornerstone of the profession, "doesn’t seem like it could ever be outsourced to an AI." This sentiment underscores a fundamental philosophical challenge: while AI can replicate appearance and voice, and even process information, it lacks the lived experience, intuition, empathy, and ethical reasoning that underpin human trust and journalistic integrity. The absence of genuine curiosity, critical thinking beyond programmed parameters, and the capacity for nuanced judgment inherent in human journalists remains a significant differentiator.
Beyond Media: Corporate Applications and Ethical Considerations
Outside the realm of journalism, the concept of digital self-cloning holds considerable appeal for corporate America. The promise of an always-on version of oneself, capable of answering questions and handling routine tasks, could revolutionize productivity. Imagine returning from a holiday without a backlog of emails or missed opportunities because a digital twin maintained a constant presence, managing inquiries and providing information. This "always-around" capability could free up human capital for more complex, creative, and strategic endeavors. Industries like customer service, technical support, and internal training are ripe for such AI integration, promising efficiency gains and improved scalability.
However, the rapid advancement of digital avatar technology is not without its ethical and psychological implications. The journalist, after the initial novelty wore off, found herself examining her avatar closely, waiting for a spontaneous blink, a new expression, or a sign that "it knows." This longing for a deeper, more autonomous interaction highlights the uncanny valley effect and the potential for human psychological entanglement with increasingly realistic AI.
The deterministic nature of the journalist’s current digital twins, confined to pre-programmed responses, prevents such spontaneous behavior. Yet, the article posits the ease with which one could "fall into a tinge of AI psychosis" with a non-deterministic avatar, powered by a chatbot free to pontificate and simulate genuine intelligence. This raises profound questions about the boundaries of human-AI interaction, the potential for emotional manipulation, and the psychological impact of engaging with entities that mimic sentience without possessing it.
Furthermore, generational attitudes towards digital twins are predicted to vary. The journalist’s observation that "my generation, Gen Z, likely won’t get used to them" stems from the "sci-fi" feel of these technologies, where cinematic fantasies are now real. This contrasts with a perceived greater acceptance of purely digital avatars compared to humanoids, suggesting a preference for digital entities that can be easily "logged off" if interactions become unsettling. The underlying concern revolves around control, authenticity, and the potential for digital entities to blur the lines between reality and simulation.
In conclusion, the creation of a journalist’s AI digital twin by Synthesia serves as a powerful case study for the profound shifts underway in professional communication. While Synthesia’s technological prowess promises enhanced efficiency and novel forms of interaction, it simultaneously compels industries like journalism and corporate affairs to grapple with fundamental questions of trust, authenticity, and the very essence of human connection in an increasingly AI-driven world. The journey into digital twinning has just begun, and its full implications are only starting to unfold.
About the Author
