ChatGPT may have taken the Internet by storm, but its interaction methods are still very limited. Chatbots of the future, however, may not be so limited. In fact, instead of being forced to type or read messages, they may be able to converse with human-like avatars.

I spoke with Israeli company D-ID about this possibility at the Mobile World Congress in Barcelona, where D-ID (opens in new tab) is committed to creating some sort of digital human, and now it is looking to add more human-like It means adding a more human connection.

The API provides real-time streaming capabilities, allowing users to talk to chatbots just as they would to regular humans. The chatbots themselves can use text-to-video conversion technology to give the impression that the digital avatar is actually speaking in a human voice. Of course, if typing is preferred to talking to a machine, it can type instead.

Being an API means that the D-ID system can be integrated into other apps and services. The general idea is for companies to build and offer virtual assistants that allow people to connect on a more personal level. [According to D-ID CEO and co-founder Gil Perry, humans are not suited for text- or voice-only conversations. Instead, having a human face (or at least a representation of one) makes everything more engaging and natural.

Additionally, from a business perspective, having an AI do a lot of the work for you can actually save you the cost of hiring someone. Whether that's talking to customers, training employees, or anything else that AI can handle in its current state.

People are currently quite skeptical and wary of AI, and Perry suggested that this faceless nature may be part of the reason. Adding a human face might make people feel more comfortable talking to a chatbot, even if they know there is no human on the other end of the phone.

I saw various D-ID chatbot demos at MWC, including the new integrated streaming capabilities. It was certainly interesting to see the various types of avatars in action, but even the most advanced ones had a bit of an uncanny valley effect.

That was especially true of the talking animations. Everything was not perfectly synchronized, and at times it seemed as if the avatar was moving its mouth randomly rather than trying to vocalize a sound.

We also noticed a distinct delay in speech recognition and writing when the bot was interacting with people. However, the fact remains that you can talk to the bot, make it understand what you are saying, and it will respond without you having to type or read anything.

The only question is when such an experience will be available to the public; D-ID's goal is to provide an API to businesses and enterprise users. However, as seen in the recent explosion of ChatGPT, there is a good chance that this type of technology will appear in a more general venue. In particular, there seemed to be a number of AI-focused companies exhibiting at MWC, and a number of ChatGPT competitors have emerged in recent weeks.