Robots Need to Learn to Keep a Secret

Robots have already staked their claim on our future. They’re taking our jobs, making art, even making love (kind of). But as we get increasingly close to our robotic pals, so they get closer to us.

In learning to better work for and with us, especially if they’re helping us out with our daily activities, they’re learning more about us too—our habits, our movements, our interests. But if recent surveillance activity is anything to go by, sharing our secrets with technology, however humanoid, isn’t necessarily a good idea.

That’s why researchers at Oxford are looking into privacy concerns surrounding robots. According to a statement on the university site today, their EPSRC-funded project 'Being There: Humans and Robots in Public Spaces’ will be the focus of this year’s Oxford London lecture later this month.

The idea is to get privacy systems written into robots before we give away too much about ourselves. Brown suggested that as robots get more human-like and we get more used to them, we might let our guard down a bit and reveal more than is necessary or advisable.

“It is important, therefore, that we design robots that have privacy embedded into their design, so their information gathering is restricted to what is needed to interact and carry out their tasks, and information about the identity of their human users is kept to a minimum,” he said. “Otherwise, these robot ‘friends’ could betray the trust of the people they come into contact with, passing on information to third parties.”

You can draw an analogy with existing technology like social networks and smartphones. Sharing information about who we are, where we live, and so on, is really convenient when you want to find friends you have something in common with or figure out the route back home, but you don’t necessarily want those details shared with third parties. If information is stored on the device—be it a phone or your robot BFF—there’s also the risk of it being hacked.

An example the researchers give of the techniques they’re developing is a method of connecting car pool partners without each individual having to actually share their home address or work route with each other.

Of course, the robots themselves aren’t to blame for security lapses; at least not while they’re still made by humans. In a video promoting the upcoming lecture, researcher Ian Brown explains it’s a matter of convincing businesses to respect user privacy from the start, not just in regard to robots but with all internet technologies.

“What I’m going to try to demonstrate in my lecture is that society and policy makers do have an option here; that we can preserve privacy in the internet age,” he said. “It’s just a question of finding a way of persuading mainly businesses that systems should be built to protect privacy—that laws should be passed to encourage that, and also to enforce privacy protection.”

The research team are going to continue their work by introducing members of the public to humanoid robot Nao next year—the same kind of model that's been hanging out with Southern Evangelicals recently. For now, regardless your faith, you might want to refrain from whispering your darkest secrets to him.