If you’ve been watching Humans on Channel 4 over the last few weeks then you’ll be aware that the Synths (life-like robots envisioned by sci-fi hungry producers) are connected to the internet. This connection, which may seem rudimentary, could have huge implications.
Much of the discussion around the dangers posed by robots of the future has focused on artificial intelligence and how super-smart robots may one day decide they don’t actually like us humans all that much and turn against us.
But few people are talking about what happens if a robot or an AI is hacked by a human that wants to cause harm. There’s been a lot of discussion around what happens when an internet connected car or an internet connected plane is targeted by a hacker but now the conversation needs to move on.
A senior researcher from Russian security giant Kaspersky told Techworld this week that it’s possible to hack any internet-connected machine if the right security protocols aren’t in place.
“Theoretically there’s no reason why it couldn’t happen,” said David Emm of Kaspersky Labs. “If it’s got connectivity then why not. It’s kind of a scary thought.
“People hack into computers and you gulp sometimes. You look at the Office of Personnel Management in the US or you look at Sony. Or you take Stuxnet and you think, ‘hang on, you’re going into a plant in another country and you’re sabotaging a physical process’. That’s pretty awesome! I wouldn’t want to say that somebody couldn’t subvert another machine.”
While Synths haven’t yet found their way into Dixons, there are armies of researchers across the world looking to make machines more intelligent. Google, for example, splashed out up to £400 million on DeepMind, a company creating artificial intelligence “agents” capable of thinking and learning independently. Then there’s Honda, which has built Asimo – a robot with a human-like body and a slightly creepy wave.
Those developing internet-enabled robots need to remember that security should be at the heart of every connected machine. If it’s overlooked then whose to say what the hacking adversaries will tell robots to do in the future. I don’t know about you, but a hacker-controlled robot wielding a gun or a knife isn’t a robot I’d like to encounter.