Let’s build AI that connects us

Original article can be found here (source): Artificial Intelligence on Medium

Let’s start with having a look at what the best possible outcome would be. To do that, we need to think about the human qualities and needs that will stand the test of time. To quote Deborah Schultz- “Technology changes, humans don’t”. It’s important to remember that the purpose of technology is that we use it to improve our lives, connect and create.

Together we’re strong

One of our most basic needs as humans is the need for connection. We all need to feel like we’re a part of a context with other human beings. Throughout history, humans have strived towards being more connected. In the early days of humanity, staying connected to the tribe was the most important thing in our lives. If we became outcasts, we were likely to die. Today this is not the case, but our need for connection remains and is likely to do so even as we’re moving on to the future. Let’s make sure that new technology like AI helps us become more connected as a society rather than drive us further apart.

What does it take for us to stay connected as a society? The internet has made it possible for us all to connect worldwide, across country borders and societal gaps, but to truly feel connected we need something more than the mere possibility. To stay connected we need to feel empathy for one another and understand our differences.

Empathy is key

What makes us feel empathy for people that are different than ourselves? The best way is to be presented to other peoples’ realities and understand their feelings. Normally this happens when we talk to other people, but what happens if we start interacting more and more with an AI that is programmed to learn about us and only present information that mirrors our own reality? We could completely opt out on getting information that isn’t directly related to us. This might be convenient, but at a large scale it could make people understand each other less and feel less connected. This is something that we need to have in mind when deciding how personalized our AI should be.

A human-like social robot made by Furhat Robotics

Space10 are trying to figure out how people want AI to work by asking them. Let’s say that this research shows that people want their AI to have the same opinions that they do. The same religion and the same political view. Even the same gender and same tone of voice. This would lead to a decrease of input from realities that are different from yours. How can people become more empathic if they have an unrealistic strong belief in their own worldview and a narrower picture of others peoples’ realities and opinions?

The danger of disconnection

So what’s the danger with this? To predict what could happen we can take a look at what large scale effects we can get when algorithms monitor our social media feed and builds what’s known as filter bubbles. In the United Kingdom these filter bubbles were used to send out targeted ads with misinformation prior to the Brexit election. A huge amount of people got false information about how many refugees UK received due to being member of the European Union, and false information about how this affected the country. The people saw this information over and over again and started to believe that it was true. This led to that people saw refugees as a threat to welfare in the UK and it created a massive disconnection between people. It created a massive feeling of “us against them”, which never leads to any good.

Example of fake news that was spread in the UK prior to the Brexit election.

This event shows how much people are affected by the information they get and how incapable they are of processing the information with a critical mind. It could become even harder to be critical towards the information given, if the voice tells you about it sounds and behaves the same way that you do.

How can we reduce the risk of people getting negatively affected and disconnected from each other by the information they get from their artificial friend? When Eli Pariser talked about filter bubbles back in 2011, he explained that the best way to curate peoples’ feed is to give them a little bit of each. Some posts that are relevant for each individual, some that are important and some that gives another point of view or that are uncomfortable and challenging. Maybe, this kind of code is what we need in our Artificial Intelligence as well. The AI that we talk to and have integrated in our lives should only adapt to us to a certain degree. It should talk to us about things that happens to other people around us and help us understand different points of view. Not only strengthen the beliefs that we already have about the world and our society.

We still need each other

Another way that we connect with each other as humans is to collaborate. This is one of our most characteristic traits as humans and it has been our way of making progression throughout history. When we listen to each other we can start to think outside the box and use our combined knowledge and experience. To ask each other for help is a very powerful tool, but what happens if we become passive and rely more on AI to find solutions to our problems? When we make decision merely based on data, we hinder our capacity to do remarkable things. This is why I think that we should be careful with what kind of solutions our AI will present to us, and to what extend it will do so. Maybe we should program it to show us new things, inspire us and then let us make our own connections and decisions.

The future of UX and AI

When designing new technologies like artificial intelligence, the role of a UX designer will be more about giving the users both what they say that they want and what they actually need. We will have to go beyond interviews and user testing and understand how to design something that serves humanity as well as the individual. I believe that the span of what we consider a user experience will broaden a lot more. When designing new technology we should look at what impact the product has in each individuals life, and maybe even in the lives of people around them. We need to see the bigger picture and understand what impact technology has in our society since we now have the possibility to have a big impact, both positive and negative. If the day comes when we all have a personal AI assistant, I hope that UX designers globally think about how this can help us become more connected instead of driving us further apart.