Tara Chklovski:What problem or area of research are you working on?
Elisabeth Lex: I’m working on a number of different projects at the moment! One of those projects is improving recommender systems. We look at how humans behave on certain tasks, and then design algorithms that are more effective and personalized. I like this field of research a lot because it’s very interdisciplinary.
Another project I’m working on looks at patterns of collaboration within networks to identify the factors that can improve collaboration, or make it more difficult. We’re also studying the effect of social status on how opinions spread over a network.
Finally, another of my research topics is open science. We want to open up the scientific process, the ivory tower of academia so all people can benefit from scientific research — not just the researchers through their publications.
TC: How are you doing that?
EL: We are diving into European policy-making. I was involved in several initiatives in the European Commission and we published documents and policies on how, for instance, scientific outreach can be measured using social media activities. We also developed a tool box to help researchers share their work with the general public. For that, we looked at successful examples of science in the media, like the Pluto fly-by, which was everywhere in the media and got people really excited. That’s a great example of how to involve people in the fruits of scientific research (which they’re also paying for).
Finally everything that we develop is open-source. We’re trying to use publicly available datasets as much as possible so that everyone can take our stuff and build upon that.
TC: What advice would you give children and parents to become more involved in understanding and using technology to improve their communities and their own environments?
EL: Think of something that really annoys you! For instance, a few years ago when I was in Austria, the government released an open data set on Vienna, and the first application that people built was an app to find public restrooms! It sounds like such a simple problem, but apparently it was a big enough problem that it was the first thing people thought of. Think about something that you have had trouble with in your daily life for a long time. It’s okay if you don’t know if there’s a solution to the problem, it’s just a starting point. The next step is to connect with others to see what their viewpoints are on that to further refine the idea.
TC: What do you think makes a good product?
EL: A good product is one where the users are involved in designing the product. Roll out your prototype to your target community, let them work with it for a certain amount of time and get their feedback. Then you really stick with them, working closely with them thinking about how to improve it – not just monitoring clicks!
TC: What inspires you?
EL: Talking to interesting people. For instance, it is an inspiration to talk to you because yours is a completely different viewpoint than mine!
Also, going to different places and learning about different cultures is very inspiring to me. Currently I’m doing research in Germany and I’m also learning a little about German culture. I also read a lot of biographies. I am reading about Elon Musk right now. It is inspiring to see how other people live their lives. I also recently read the biography of Chris Hadfield, the Canadian astronaut who was the commander of the space station.
TC: I love that book! It is such an amazing story of a childhood dream that impossibly comes true. He is a little kid on a farm who makes up his mind to be an astronaut, works towards it and becomes one! What do you find difficult?
EL: I find many things difficult! In terms of AI, I am troubled that a lot of the efforts out there are optimized for more clicks and views. I’m not only interested in large reach, as this also has harmful consequences. For instance, I am very troubled about the issue with videos targeted at children in YouTube videos . There is a lot of malicious content now that is being injected into children’s YouTube videos. You could see Peppa Pig, or another children’s character killing another character. The hypothesis is that because children’s videos get a lot of clicks they make a good vehicle to carry your message to many people. It’s not completely clear who is doing this and why. But this is stuff that troubles me. It’s also difficult to figure out how to detect and stop this sort of thing.
TC:Right. This is the age-old battle between tools being used for good and bad purposes!
EL: Another thing I find difficult is that my goal through my research is to help people, which is sometimes in opposition to what companies want, which is to become richer. Both groups don’t always want the same thing. When I build a decision‑making system for people I want to give them access to the best content, which may not be the most popular content!
TC: How do you think AI can help strengthen societies and communities?
EL: To a certain extent, it already does that because it lets us connect even though we’re in completely different countries at different time zones, and we have never met before! I’m pretty sure that AI will create a lot of new opportunities, new jobs and new tools that we won’t even be able to predict. My real hope is that AI allows humans to shift their focus to more interesting stuff, by automating many of the repetitive tasks. It will allow us to do something completely out‑of‑the‑box.
TC: What field of AI excites you the most?
EL: This is so hard to say! I really like the idea of autonomous cars. And then the use of AI in health care is also very exciting — in diagnosis, but also in training doctors. I was visiting a colleague last week and he showed me an augmented reality model of the brain that you can rotate and observe how it really works without cutting somebody open.
In the field of learning, there’s a lot of potential to better target students based on their needs in order to predict what kind of support and training they need.
TC: So, last question: how do you think children can learn more about AI?
EL: I think this is tricky because it’s not enough to just buy a robot or intelligent toy; although it is worth it. You need to teach your kids to be creative and curious and to ask the right questions, but I think it’s also good to show them how AI works in real examples. Show them what happens when they click on a Netflix video and what implication that has.
And at the same time, I think the parents need to learn as well! Because the world is changing — and it is exciting to be part of it!