Humans and virtual avatars can now interact with each other.
Crédits photos : Pexels / cottonbro studio
Artificial intelligence is impossible to ignore. AI is everywhere, in our personal and professional life. Without us noticing, many everyday tools are already using it, like voice assistants. Companies even train their employees to use solutions such as ChatGPT. In the immersive field, artificial intelligence changed the game and high-performance software and hardware is developing. Today, AI-driven avatars are becoming inseparable from immersive experiences, to the point of redefining interaction models. But are their limits? Find all the answers you are looking for at the conference session “AI-driven Virtual Characters and Avatars” on Thursday April, 10 during Laval Virtual 2025.
Virtual avatars: less automation, more realism
If we look back three years, the tools to create ultra realistic avatars were fewer and most of all complex, using advanced motion capture techniques. Some startups were beginning to use artificial intelligence to make avatars speak. But we were a long way from the progress made the last few years and from the quality of the current solutions. Today, generative AI has changed all that.
When we used to talk about avatars, we thought of the characters we play as in video games or virtual worlds. Now, they are virtual humans capable of understanding us, interacting with us and showing emotions. The level of realism today was never reached before. As demonstrated by NVIDIA ACE technology that creates virtual humans able to understand us and interact with us in video games.
That’s why we need to use generative AI. Compared to its sister, generative AI reproduces human cognitive functions. Traditional artificial intelligence is conditioned to perform a well-defined task, without any notion of creativity or interaction.
From now on, AI-driven avatars are capable of having a natural conversation with a human being, but also showing emotions. A content creator recently proved that using the Minimax tool to generate a video with an avatar showing disgust, joy, anger, etc. Virtual avatars have reached a superior level and we use them in every business area.
In 2023, Laval Virtual was already showcasing these pioneer uses of generative artificial intelligence. The Quantum Bar project, presented in the #Experiences competition, offered to interact with a AI-drive barman in a digital environment. We were then at the dawn of multiple uses, and virtual avatars have now reached a new level.
When I grow up, I wanna be a virtual avatar
Four years ago, we wrote on this blog about the rise of virtual influencers. Today, uses have increased. Virtual avatars can be salesmen, trainers, assistants and even healthcare agents.
Improving customer experience
In marketing first, virtual avatars were a little revolution in the world of client support. With artificial intelligence you can have virtual assistants available 24 hours a day and 7 days a week. Virtual avatars became a promotion tool for brands, which use them to create an unforgettable customer experience.
One example is the promising startup Jumbo Mana, which recreates historical characters in virtual avatars. With the solution, a person can chat with an AI-driven avatar. It’s used in the cultural field, as well as for events and retail, in order to improve the customer experience.
There is also the AskMona project, a conversational tour guide to whom you can ask questions while visiting a museum. The Musée du Louvre used this solution to engage young audiences during a temporary exhibition.
Avatars as assistants
Virtual avatars are becoming increasingly popular in training and education. They can become private teachers, to help students review their lessons. Moreover, thanks to generative AI and the evolution of the techniques, those instructor avatars are able to give personalized classes tailored to each student.
In the business world, AI-driven avatars are interesting to train in risky fields such as aviation and defense. According to a study, the use of a virtual avatar in training videos increases learner motivation and knowledge retention.
Startup Hippocratic AI went one step further and created virtual health agents. They can make check-ups before or after a surgery or follow a patient on a daily basis. Worldwide, it is estimated that there is a shortfall of 15 million healthcare workers to meet the demand for care.
Hippocratic AI’s ambition is to facilitate access to care and reduce the delays in treatment that can put patients at risk. Having avatars as health agents is quite new, and the use of AI in a domain such as healthcare raises ethical questions.
Ethical issues around virtual avatars
What about the sentence: “artificial intelligence will replace humans”? It has become so easy to create an image or generate a text. We can even have a conversation with an AI! With the massive development of generative AI, it’s fair to ask questions. Especially when we know the foundation of this new artificial intelligence, which is more creative and performs tasks that humans are capable of doing themselves. Of course, not all uses of AI are reprehensible, even if some do provoke a reaction.
We talk more and more about deadbots, these virtual avatars that bring the dead back to life. They use the personal data of a deceased person to become its virtual copy. This means, someone can “talk” with a missing relative. Of course, it’s an entirely artificial construction, but grief doesn’t always enable us to clearly see the ethical limits we must respect.
This practice can have significant psychological repercussions on users. It also raises the question of the right to use personal data after the death of a person. The massive use of Internet users’ data is also the subject of debate. It’s also an important issue for companies for confidentiality reasons. Is our data protected when we use virtual avatars?
The Council of European Union published in June 2023 the first regulation text on artificial intelligence: the AI Act. The objective of the decree is to guarantee the protection of private data when using an AI system. Its purpose is not to restrict the use of artificial intelligence, but to provide a better control structure, with the ultimate aim of encouraging innovation.
It is also the mission of the Laval Virtual exhibition throughout its conference programme: better understand and manage immersive technology. Follow the session dedicated to the uses and perspectives of virtual avatars of Thursday April, 10.