When you hear the word “machine,” what comes to your mind? Wires? Metal? A monotone voice? Pink Floyd’s 1975 hit “Welcome to the Machine?” Whatever it is, rid your mind of it before you read further. Let me tell you what a machine basically is: a man-made device created solely to perform a certain task. The same goes for artificial intelligence - it’s made to perform the task you tell it to do. Don’t worry, we don’t live in the Terminator universe where AI has become self-aware and can operate without a human giving the orders! Also, I wasn’t a fan of Genisys. Was anyone, really?
At PRSONAS™, we don’t make boring machines. Anyone can create one, but we aren’t just anyone (you know that)! Using the latest technology from our talented team, we craft digital personalities that can actively learn and adapt. Personalities are not something new to us - some of our earliest models include a chef, the Mariner Moose for the Seattle Mariners, Einstein in London, and just so many more.
If you want an example, think about Flo from the Progressive Insurance company: she’s one of the most well-known character personalities in commercials today because she’s not a lifeless being you can’t identify with; she's a personality and quite a believable one at that. Ultimately, we tout ourselves as leaders in the creation of digital personalities as opposed to mechanical spokespersons.
Now that you’ve distanced yourself from the idea of us simply making machines, let’s talk about how our personalities adapt. No... rather how they learn. PRSONAS™ units do not use machine learning to adapt and learn, they use something called “active learning.” What’s the difference between the two terms?
- Machine learning (ML) = the machine reviews a ton of data and determines its own modifications for improvement. You tell it the outcome and it determines the algorithm to get there. Sounds creepy and cold, doesn’t it? According to Investopedia:
Machine learning is an area of artificial intelligence (AI) with a concept that a computer program can learn and adapt to new data without human intervention.
- Active learning (AL) = the machine reviews data, suggests edits, and asks for confirmation from its respective human to verify the accuracy of this new knowledge. There is an active participant in the process. We can’t let the machines do whatever they want (talking to you, Arnold Schwarzenegger!) so we need to have control.
Active learning allows an AI program to identify what data it needs to become better with the help of a human speeding up the learning process.
So, what’s the big deal? Why does it matter? Who cares? The reason it matters is that this allows PRSONAS™ personalities to provide a truly human response! There are no machine responses and certainly no fear from a rise of the machines.
To aid in this process, PRSONAS™ units use Microsoft Azure’s Luis for language understanding. Now, how would this work? Here’s an example: say I walk up to Daphne and say, “what’s up dawg?” A bit of an outdated greeting, sure, but some people still use it. How would the machine respond? Nothing relevant if the computer was instructed to simply greet someone. A machine with AL will make its best guess and say “I think this is what you are trying to say” whereas a machine with ML may not even respond at all. Because of that follow-up in AL, this is the human way of asking and thus human engagement. This is how Daphne learns! In fact, she may even reply “not much, my homie.”
How Daphne responds and learns with an active human participant is all a part of the training cycle and how she can be scaled. Over time, she adapts and can understand more requests and respond in a myriad of new ways!
So, in summary. Would you rather have this...?
Or this?
The choice is yours. At PRSONAS™, we are teaching our units how to act like a human, not think like a machine. The power of personality!