Read our COVID-19 research and news.

Technology continues to pervade our lives.

Michel Royon/Wikimedia Commons

When will I have my sidekick robot?

BOSTON—From Netflix recommendations to credit card fraud detection, artificial intelligence (AI) is already part of our daily lives. But as AI expands, where do we draw the line on how intimate we become with this new technology? 
Swarm engineer Sabine Hauert of the University of Bristol in the United Kingdom is part of a Royal Society working group asking just that. Hauert, a swarm engineer who works with nanoparticles, has spent time speaking with members of the public about fears and hopes for advancing artificial intelligence. Here on Saturday at the annual meeting of AAAS, which publishes Science, she gave a talk titled, "AI and Policy Engagement: Understanding the Public's Views of Social Risk". Hauert sat down with Science to discuss the issue. This interview has been edited for clarity and length. 
Q: You’ve found that only 9% of people in the United Kingdom have heard of machine learning. But everyone has heard of AI. How do they relate?
A: AI is an abstract concept—different people have different definitions. Very often what people think when they say artificial intelligence is humanlike intelligence. Machine learning is a concrete process that is really the science of computers learning from data. We might be looking at one specific task with one specific set of data and be able to come up with a prediction or a solution based on that. And were using that as a starting point so that we don't get lost in all the discussions about what is AI and what does this technology do. 
Q: Where do we see machine learning in our lives?
A: There are examples of machine learning all around us. We see it in our spam filters, recommendations online—whether it's movies or with things that we'd like to purchase—and we see it in credit card fraud detection. And there are a number of areas where we are going to see more machine learning in the future.

Swarm engineer Sabine Hauert 

Sabine Hauert

Q: What are the goals of your Royal Society working group?

A: They’re creating a report looking at the potential for machine learning in the next 5 to 10 years, and also the barriers to achieving that potential. They're engaging with a number of stakeholders across the U.K. who'd be interested in this technology whether its industry, policymakers, academia, or the public. And they're trying to look at it from a number of perspectives: ethical, legal, scientific, and societal. 

What I love about the project is that actually a big chunk of this working group’s role is to engage with the public. We surveyed people across the U.K. and asked them what they think of machine learning. We've also had focus groups where we spend more time with small groups of people to dig in and understand what they want from this technology. 

Q: And what are the responses from members of the public you’ve worked with?

A: It's very much context dependent. People won't feel the same way if you're talking about autonomous cars versus something that can help doctors do better diagnostics. When they do see areas that benefit them, there's genuine excitement about the technology. People are worried about making sure algorithms can work with humans. They want to make sure the algorithms are safe and trustworthy. And there is the discussion about robots replacing human jobs.

Q: And how do we move forward in this field without replacing humans?

A: Well it’s about tasks, not jobs, in terms of the way that we're building the future. We now have algorithms that can detect markers of cancer in images. But the goal is to create tools for the doctors rather than replacing them. 

Q: What's your favorite example of AI in science fiction?

A: The movie Robot and Frank. It's the story of an elderly person who gets a caregiving robot for the home. He convinces the robot that to be happy he needs the robot as a sidekick to become a robber. It’s just a really nice story of the limitations of the technology, in that the person quickly understands how he can manipulate it, but also of a partnership. And even though the motivation is dubious, in the end these two end up as a genuine team.

Q: So when will I have my sidekick robot?

A: I think you'll have different technologies for different tasks, just like you have lots of apps on your phone. I'm guessing in the future we're going to get more and more of these helpers that are really focusing on a specific area. Having a fully functional system that can do everything is just so far away. 

Check out our full coverage of AAAS 2017.