./uploads/advanced-cache.php AI-Generated Humans Get an Emotional Upgrade with Synthesia’s Expressive Avatars

Logged-out Icon

AI-Generated Humans Get an Emotional Upgrade with Synthesia’s Expressive Avatars


Synthesia, a British AI unicorn, has launched a new generation of AI-generated humans called “Expressive Avatars.” These digital beings are designed to deliver the most realistic emotional expressions in the market, blurring the line between virtual and real.

The avatars are created using an AI model trained on footage of real actors. Users can simply input a text prompt, and the synthetic humans will read it out on screen, complete with appropriate tone of voice, body language, and facial expressions. This is made possible through a technique called “automatic sentiment prediction,” which allows the AI models to infer emotions within the text and adjust the avatar’s performance accordingly.

According to Victor Riparbe, Synthesia’s CEO and co-founder, these avatars are the first iteration capable of expressing emotions and understanding the sentiment of the content. The avatars also generate unique outputs for each script, ensuring that no two performances are the same.

During a product demo, TNW had the opportunity to review the avatars’ acting abilities. While the photorealistic faces, emotional gestures, expressive voices, and synchronized movements bring a new level of realism to the market, the avatars still fall short of escaping the uncanny valley. They tend to slightly exaggerate their emotions and are limited to head, face, and shoulder movements.

Despite these limitations, Synthesia’s avatars have the potential to unlock new applications. The company already has 55,000 businesses as customers, including half of the Fortune 100, who use the platform for training, presentations, marketing, and customer service. With the introduction of Expressive Avatars, Synthesia plans to expand into promotional videos, customer support, and healthcare.

However, the increased realism of these avatars also raises concerns about their potential for spreading disinformation. Synthesia’s tech has previously been used to produce fake news in China, Venezuela, and Mali. In response, the company has implemented various measures, including banning accounts, introducing new rules, upgrading digital defenses, and dedicating over 10% of the company to trust and safety-related work.

Looking to the future, Synthesia envisions AI avatars playing a significant role in the workplace, roaming 3D offices, communicating with virtual humans, and assisting with various tasks. In the near term, Riparbe has set an internal goal to produce an Apple Keynote-style presentation featuring an avatar walking and talking in a way that feels and looks entirely real by the end of the year.

As AI-generated humans continue to evolve and become more sophisticated, it will be crucial for companies like Synthesia to strike a balance between pushing the boundaries of technology and ensuring responsible use and development of these powerful tools.

Posts you may like

This website uses cookies to ensure you get the best experience on our website