I recall when the Amazon Echo arrived, and I was setting it up for the first time. When you power the device on, it informs you that “Alexa listens and adjusts to your voice over time”. This onboarding experience is personalized, and I felt a warm responsibility to help a digital assistant get familiar to my world.
In the news, the sci-fi treatment is given to AI. You know, the imagery of “Terminator” and Arnold Schwarznegger. That’s not exactly how consumers and even businesses want to see AI panning out. Digital assistants need you. They’re designed to make your life easier, after all. Firstly, they will need to get smart enough to learn a bit about you.
How does a Digital Assistant Learn?
Digital assistants vary in ability, and are often designed around a certain type of use. Immediately, we can think of Apple’s Siri, or Microsoft Cortana. There’s also the digital assistants within the Chatnels app, which can be trained to suit the type of need or conversation.
They use advanced digital intelligence, natural language processing to create a conversational experience.
From a ScienceMag article, we can start with the context of how humans learn.
“Humans are not blank slates, nor are we hardwired. Instead, the evidence suggests we have predispositions that help us learn and reason about the world. Nature doesn’t give us a library of skills, just the scaffolding to build one.”
This is different from the way we think of how humans learn, as a baby. There’s usually a mix of instinct and learning. Digital assistants on the other hand use machine learning AI to identify patterns of behaviour in data. Later, the model is optimized as more information is added. Where does this information come from though?
Digital Assistants Learn… from Us!
For the most part, digital assistants will combine historical information such as our search history, location, demographic, or even purchase preferences. Then, algorithms create data models that then feed into a digital assistant’s miraculous abilities to recommend you a song, or make predictions for what you will say
Oracle has an article on Digital Assistants that goes in depth with explaining what they are, and the type of business functions they can perform. It includes automating responses like an help-desk, to transferring users to the right staff. Sometimes it’s seamless, where a chat can transition from chatbots to a live user on the other side.
We don’t build friendships with digital assistants. Undoubtedly, we provide the data to improve their jobs and are depended on to do so. And their jobs are to help us.
Here’s one of my all-time favourite articles on a digital assistant. Its name is Jibo that said goodbye to its users. The tale is bittersweet, and it reminds us that the simple relationship with our digital assistants can be interesting food for thought.