As someone grown up in the 90s, I’ve always been fascinated by robots. The idea that, at some point in the future, we would be able to teach them how to take care of the most boring tasks was desirable, the possibility that they might replace us in various professions exciting and a little scary.
Twenty years later, the future is here and yes, it is offering us artificial intelligence. We own voice-based assistants helping us take care of the house and smart vacuum cleaners that learn the planimetry of our flats and clean them for us.
Self driven cars are a thing. It’s exciting. However, like every other new big technological disruption, AI brings both trepidation and curiosity, and a rose of doubts, fears and confusion. Everyone knows how crucial user experience is in human-computer interaction, but how do we approach this completely new way to communicate with machines?
Which are the biggest challenges for the UX community?
1. Build trust
Trust is fundamental for every kind of successful relationship.
I am able to trust my co-worker even if she has sent me the wrong image for a presentation. Why? Because, as a human, I understand other humans’ nature to make mistakes. I empathize with it.
The situation is different when I deal with machines and systems. In this case, I demand a different level of efficiency, I require an error-free practice. The paradox is sort of funny: computers avoid mistakes humans often make, but fall into traps a real person would easily obviate.
One of the biggest UX challenges is to help the user understand the fact that products related to AI aren’t infallible yet. What if this just makes machines a little more “human” and thus more similar to us?
2. Improve communication
In order to achieve a pleasant and natural user experience, users should never perceive AI as an hostile entity. A bot that misunderstands commands or that depends on non-intuitive layers to complete a task will always create frustration.
In order to become a positive company, a friendly and natural tone of the communication should be preferred. In addition to the register, we have to remember how we can guide the user towards the achievement of his or her primary goal as fast and as easy as possible.
Does the conversation feel like escaping from an intricate maze? Does the user waste more time interacting with the machine than completing the task him-/herself? Does the user experience cause confusion because the system doesn’t recognize the way they talk? Maybe, instead of trying to copy human language, we should be able to create a new one, specifically for this new type of communication.
3. Increase customization
Popular wisdom says that “It takes all sorts to make the world.”
We are all different individuals. Something that is irrelevant for me can mean the world to you. AI should look and feel like it knows its users well and behave accordingly.
Do you like it when Alexa tells you sassy jokes or do you prefer it to just quickly read your daily to-dos? The system should be able to adapt to its users and become more and more helpful over time. To do so, the user should have the option to provide feedback.
A deeper, more personalized level of testing is also important to grow confidence in the user. If done correctly, the users can finally start believing their AI is a product made to improve their lives and a friend that is here to stay.