The Real Future of Chatbots Is Less About “Smarts” and More About “Feels”
The AI behind chatbots and personal assistants like Siri and Alexa have come a long way in a matter of years. It’s scary how effective they have become at getting a lot of things done for us, be it checking the weather, booking tickets, finding stuff online, and more.
But all these assistants and chatbots have one thing in common: they are often too cold and professional, a bit too robotic if you will. Sure, they can come up with snarky and funny responses when you prod them a bit, but ultimately it is a shallow experience.
The best these chatbots can do is deliver information, or carry out simple tasks. While they can satisfy basic queries, they cannot make you feel better.
Looking For Something Deeper
For that, you need a bot capable of figuring out how you are feeling at any given time, based on the way you respond to it. If you have watched the Joaquin Phoenix movie “Her,” you will understand the difference. If you haven’t, give it a shot. It’s worth it for Scarlett Johansson’s voice alone!
Chatbot AIs are trained to look for specific phrases and words to work up a meaningful conversation. But they are blind to the context and emotional weight of the words we type, or say, in the case of Voice Assistants.
Deep learning and neural networks are what you need to make chatbot AI smarter and more sensitive. For example, researchers at Tsinghua University created a chatbot capable of identifying the emotional context of specific Chinese phrases in a conversation. They used advanced algorithms, big data, and deep learning to create this bot.
Why Do We Need Emotional Chatbots?
There are numerous chatbots out there these days with some level of emotional capabilities. In most instances, these are specialized software with a quite narrow purpose.
Take the Woebot for instance. It is a therapist chatbot trained to talk to mental health patients on a regular basis. Most of us have some inhibitions when it comes to speaking frankly to another human about sensitive personal matters. Knowing that an AI is listening can actually help people open up and seek the help they need.
Spot is yet another example of a sensitive bot that feels like a perfect fit in today’s #MeToo zeitgeist. Using open-ended questions from police interview techniques, the chatbot encourages users to report instances of sexual harassment and other troubling behavior at the workplace.
In business, emotional intelligence can be invaluable, providing customers with a richer, more satisfying experience than ever before. And you don’t necessarily need a product or service to push. Just providing people with a bot they can have an engaging chat with could pay off big time in the long run.
This is why all the major players like Google, Facebook and Apple are busy working on empathy-driven chatbots for their messaging platforms. Even Microsoft seems to have recovered from the horrifying faux pas that was Tay, back in 2016. It’s successor Zo is predictably very politically correct, but seems to have a few neat tricks up her sleeve.
The story of Eugenia Kyudo, who built an AI chatbot based on the personality of her dead friend Roman Mazurenko, is poignant and disturbing in equal measure. Designed using neural networks, deep learning, and data gleaned from Mazurenko’s chats with family and friends, the AI chatbot was capable of feeling like Roman, even to his family and friends.
So if you ever feel like attaining a form of limited immortality, a la Black Mirror ”Be Right Back,” you should consider adding this provision to your will!