Conversational AI is a very young field – although some might rightfully point out that it really has been around ever since Alan Turing proposed the Turing Test. That said, it hasn’t been until the past few years that we’ve started really seeing adoption of technology that we interact with by talking with it.
For example, my friend just asked me if I had Amazon Prime because there’s a hot Prime Day deal for a $5 smart plug with a catch – you have to order it by asking Alexa, “Alexa, order the Amazon smart plug.” I tried it out and turns out it was a sold out deal, which means that a whole lot of people must have asked Alexa to order it. The point is that we are starting to see more widespread adoption of voice enabled digital assistants like Alexa.
But if we peer into the future of conversational AI, what comes next after chatbots and voice assistants? We propose the concept of an AI avatar that includes the functionality of chatbots and voice assistants and extends them in the following three ways:
AI Avatars are animated using deep learning to bring the AI character to life in a way that we can see on the screen. Matching a face to the voice of an AI takes the experience of talking to an AI to a whole new level! This is because communication for humans is not only verbal. A lot can be communicated with just a glance, so much more than just the words that are spoken. As a matter of fact, we have entire regions of the brain dedicated to recognizing faces and forming more complex perceptions of others based on their facial expressions. When we are able to see the AI avatars it creates a much deeper connection than only hearing them.
Chatbots and voice assistants are fairly utilitarian. While they may get the job done, you don’t usually feel like you’re having a conversation. They’ll help you book a hotel, set an alarm, or play games with you, but their personalities are limited by their given task and the fact that they are designed for a large, general audience. AI avatars are not trying to be everything to everyone, but rather they are very specialized AIs designed with a specific audience in mind. They are specialized in specific niches with personality traits that resonate with their users and represent a persona that the user would want to talk to about a specific topic. With this kind of presence comes a higher measure of cognitive authority in AI avatars. For example, if you were to ask both Alexa and a pharmacist AI avatar for specific details about a medication you were recently prescribed, you would probably trust the pharmacist AI avatar’s opinion over Alexa’s, for the simple reason that they are perceived as having more knowledge in the field.
An AI avatar creates memories about its users so that future conversations don’t have to completely start over from scratch. This is a concept known as persistence and is unique to each user. Having a memory allows the AI avatar to remember what you may have told it the day before, or the week before. This differentiates from chatbots because these simply store your information as just that, information. The AI avatar can also cater to your specific needs. It’ll learn what you like based on what you tell it you like, as opposed to chatbots who won’t know the difference between you trying something new and hating it or liking it, yet making suggestions based off that. When you talk with an AI avatar that remembers what you previously spoke about it creates a stronger connection than what you would get from a programmed bot just running off of a script.
These are just three of the ways we see conversational AI transforming in the coming years. Keep an eye out for more articles in the future were we explore concepts for a variety of niche use cases for AI avatars!