What I’m listening to as I type: The Rise And Fall Of Ziggy Stardust And The Spiders From Mars (30th Anniversary 2CD Edition)
Among the umpteen things robots struggle to do are climb stairs and show empathy.
CoBot, developed at Carnegie Mellon university’s Robotics Institute, deals with the stairs issue by waiting by the lift and begging passersby to press the button for the floor it needs.
In asking for our help, the robot is tapping into the empathy most of us feel when we see weakness in another. Even an inanimate other. (Although I suspect more than one mischievous human will have sent CoBot to the wrong floor).
I’ve been thinking about robots recently. Partly because of an excellent report in the Economist and partly because I’ve been thinking about what journalistic skills can and can’t be taught to the drones and algorithms doing some of our reporting.
I once worked with a young reporter who was a dreadful writer and a promising journalist. When the newsroom emptied into the local pub at the end of a long shift, he’d work the bar, glass of coke in hand (he wouldn’t drink), and chat to everyone – hunting out stories and gossip and contacts.
While we sat in “our” corner, moaning about editors and readers and the unrealistic expectations of both, he would still be working. I once asked him why and he was confused by the question. “Why wouldn’t I?” he said.
Nosiness; curiosity; a need to know what others know; always wanting to ask the question beyond the comfort zone; being the stranger people open up to in a bar, on a train, at a crime scene… are these journalistic skills or personal attributes? And can they be taught?
I tell my students that the most important skill a journalist can have is curiosity.
Not only about the big stuff – who owns what, why that war started, who’s spying on who – but curiosity about the everyday. What’s going on with that couple having the whispered argument? What brought that guy to counting up pennies to buy his pint?
News is what’s happening, journalism is what might be happening.
The stories that young reporter picked up were sometimes great, sometimes dull and usually needed rewriting. Narrative Science’s news-writing algorithm would probably have done a tidier job.
But robots and algorithms can only work with what we give them. They need us to push the button to get them to right floor. And no matter how many buttons we press, we can’t make a robot curious.
(I realise I’m concentrating on words here, but ask any photo-journalist whether their best picture is down to gut response to a moment or an eye for framing a shot. We can only teach part of both.)
Let’s wander back to that pub and imagine my junior reporter being replaced by a story-seeking robot. Say 10 years from now (because the stair climbing thing is still an issue).
Robots don’t do thinking on their feet. They need to have a pretty basic mission – like hoovering up crumbs, mowing grass, or pulling hospital trolleys, to be able to do the job without needing a human operator. Variations in environment or task usually need to be dealt with by programming on the go.
Our story-seeking bot would have sensors so it could maneouver around the bar without bumping into people or knocking over drinks. It could be programmed with a databank of names and faces so it can spot the local Councillor or retired footballer. All of these are possible now.
But it’s the empathy bits that our bot would struggle with. How does a robot know when to make eye contact and when to look away? How does it know when it’s better to change the subject? When to press for a response and when to let someone be? When to ask a different question to get a better answer?
Asking questions – it’s not a game
Maybe our story-gathering robot doesn’t need empathy, it just needs to listen in a non-threatening way.
Alexander Reben and Brent Hoff are filmmakers using story-gathering robots to document how and why we would open our hearts to robots. Their Cubie bots are small and cute with smiley faces drawn onto their cardboard shell.
They’ve been programmed to ask psychologically-proven questions to encourage people to open up to them.
There are things robot reporters are already doing pretty well – like providing cheaper disaster footage; or data-reliant stories for tiny audiences. These are not robots as we imagine them, but software and drones.
Media outlets are increasingly using algorithmic software to produce stories and drones to capture eyewitness content. Data is fed in, the algorithm adds structure, and an acceptable story is produced. Narrative Science’s algorithms fill the reporting gaps left by our contracting news industry – school sports, business data, etc.
It’s all our fault really – us teachers of journalists. Every time we teach a trainee to write a story with the who, what, where and when within the first two pars, we’re making it easier to write by algorithm.
Two years ago, scenes of flooding in Wiltshire were captured by a viewer with his own drone and sent to the BBC. Now every man and his newsroom seems to own a drone. Here’s the Telegraph’s drone report from the floods earlier this year, and flood footage posted in the last few days by the Washington Post.
In the US, media companies recently filed court documents arguing the Civil Aviation Authority is “hindering” free speech and press freedom by restricting commercial use of drones (ie, use by newspapers but not by government agencies or hobbyists).
But with 954 journalists killed in the last ten years (to time of writing), how long before we’re habitually sending robots into war zones instead of reporters?
A robot can capture pictures, record what people say and be controlled at a distance by a journalist-operator, perhaps feeding it questions to ask. Safer, remote gathering of what’s happening on the ground could mean we get closer to the reality of those trapped by war or disaster.
With my ex-news editor head on, I can see the usefulness of robots, drones and algorithms as reporting tools.
I doubt I’d ever send a robot to newsgather in a pub (unless it was full of zombies or bombs) but I could see me sending it to report court proceedings, or council meetings, or a press conference – anywhere where being there to gather details, a photo and a couple of quotes is better than not sending a reporter at all.
We shouldn’t see these as technologies that will take jobs from journalists (that’s happening in any case) but as tools that could help us report more of what’s happening in the world – and perhaps push journalism to concentrate a bit more on the ‘why’, rather than the ‘what’ or ‘who’.
Robots are only what we make them to be. It’s our choice whether a drone is an eyewitness, a spy, or a killer. And the line between eyewitness and spy for drone journalism is one we’ve already crossed.
The key question is not what can robots, drones and algorithms do now, but what might someone might make them do in the future? And that includes how we choose to use them as journalists.