We need to talk, computer!
Artificial intelligence in conversational interfaces.
Oh, what an amazing time we live in! We’re talking about colonizing mars, curing any diseases, living eternally and what we can do, to avoid a war against an intelligent robot species that we actually built by our own.
I agree with you, that some of the examples still sound like Sci-Fi, but especially the “intelligent machine” isn’t that unrealistic as you might think. No worries, there’s still a long way to go until Schwarzenegger knocks on your door, but there are already highly artificial intelligent systems in use in the industry and also available for everyone today. This leads to the question, what does it mean if a system, machine, computer or robot is intelligent?
What is intelligence?
That’s a very controversial question! So let’s start with a try of a definition from Wikipedia:
“Intelligence has been defined in many different ways including as one’s capacity for logic, understanding, self-awareness, learning, emotional knowledge, planning, creativity and problem solving.”
This controversy is a big issue in developing AIs. If we can’t even tell exactly what intelligence is, how should we teach a system about it?
Since the days of Alan Turing until today, a regular computer-system needs extremely precise instructions to solve a task as expected. It also needs descriptions how to achieve the solution for the task in every detail. So, without humans, it’s just a very stupid bunch of electronics!
Today’s artificial intelligence fails in almost every point of the definition as well. They are not able to grow an emotional knowledge, are not creative and have no self-awareness. There’s just one big difference compared to a classical computer system: Learning!
You are a teacher, deal with it.
A modern AI behaves in many ways like a human child. It learns from observing, listening, asking, interacting, failing and succeeding. The more information it gathers, the smarter it gets.
Like it or not, probably every bigger digital service that you use collects your data and behavior to learn from it. If you feed an artificial intelligence with tons of data like this, the system starts gaining a knowledge of the world it’s living in. This world could be a Zalando website, or it could be a network with billions of information, fragments, pieces etc. like the Google Search.
If you start interacting (let’s say chat) with an AI that has this knowledge, you might get an impression of smartness. But is it intelligent? Based on its knowledge, an AI can understand your questions and might give you logical, problem solving answers. It could even recognize that you are sad based on fluctuations in your voice or face recognition and all this fancy stuff, but it just won’t get it WHY you are sad! It is simply not able to anticipate the missing information.
Even if you explain why, the system will save this information like any other. And that’s why we are still far away from an AI like we see it in Sci-Fi movies, far away from a system with self-awareness which fells emotional, subjective, intuitive or creative decisions and far away from a human-like artificial intelligence.
Now after I dissed AIs a lot, let’s be fair and take a look on the advantages of these systems, what are they able to do today and in the near future and how we can use it in a meaningful way?
Conversational Design FTW!
It’s pretty obvious that the field of tech, science and research owns the most capable artificial intelligent systems nowadays. There are Jeopardy playing supercomputers like the IBM Watson or PRISM, an NSA program which conducts data mining and filters content self-sufficient.
Besides the Google Search, no artificial intelligence is appearing prominent in our everyday lives. Even if the technology is getting cheaper and evolves rapidly, it’s still at the beginning. We all probably used “Siri” once or twice because of it’s funny, pre-programmed answers, but it isn’t a useful tool yet.
AI driven Chatbots are the latest hype for e-commerce and customer support solutions. There are a lot of advantages like you can lead the user much better in a simple, reduced chat interface, instead of displaying huge navigations and lists. It also feels much more natural if a Chatbot tells me that it couldn’t find an information, instead of displaying a 404 page on a website.
Conversational interfaces can also enhance the user experience based on every single users needs! Individual, non-linear user flows are the absolute superpower of this technology in term of Interaction Design. Imagine an interface that is able to interact differently based on cultural, demographic and even sociological aspects of each user. Or an interface that recognizes the users current abilities to interact with it (e.g. driving on the bike and looking for a certain address vs. sitting with headphones in front of a desktop computer) and offers the best fitting interface for this specific current situation.
So it’s more than just a text-based input-output interaction. It will change our mindset of how an interface should look and behave. Powered by artificial intelligence, it’s the beginning of achieving a real conversation and a better communication between humans and machines.
But is there more?
What else can we do with AI driven Chatbots beside some interface enhancement on Amazon or make some small-talk with a cat called “Poncho” about the weather and stuff?
What about using this technology in term of sustainability, accessibility or to save resources? If you’re curious about this question like I am, you should follow me here on medium or on twitter.
And you should join my talk with Designers and AI-Experts about this topic on April 20th, 2017 in Bern, Switzerland. Just ask for further details.
…or just leave a comment if you wanna chat with a real human ;)
