Newsroom of the New York Times, 1942
Last week I outlined some of the interesting things that are going on in journalism with the increasing use of algorithms to mine data, detect interesting trends and even to write entire articles. At the moment there isn’t much discussion or anxiety about this trend, perhaps partly because the technology is so new to our field and past examples suggests it’ll take a while to make itself really felt. It took the best part of two decades for algorithmic trading to really take off in finance, one of the few areas to date where these technologies have really transformed working practices, but now the industry has been completely changed by it. As I noted last week journalism might be next, but what about photojournalism?
James Kotecki of Automated Insights suggests that those nervous about an algorithmic future should ‘get a job they can’t automate’. Given the nature of our work as photographers we might think that we’re in a pretty good position in that regard, and certainly for now we are probably the furthest from being impacted by these sorts of technologies. Algorithms are best suited to making use of certain types of data, and photographs (and video) are still notoriously difficult things for algorithms to deal with and extract useful insights from. Google’s attempts to teach algorithms to caption photographs demonstrates this. What they’ve achieved is pretty remarkable, but it’s still far from what a human can do. Still, the precedent is there so perhaps photo editors if no one else should beware.
As photographers we also might feel pretty safe knowing that in the end our job involves physical as well as mental work, it involves going out into the world, encountering and interacting with people, telling stories, all things that it’s hard to imagine an algorithm taking away from us. All true, and all reassuring, but also rooted in an old model of what photojournalism is, and this might be the major thing we’re overlooking.
What so-called disruptive technologies like algorithms do so effectively is to destroy those old ways of working and thinking and something, and usher in new ones. Algorithms might struggle to take over many of the tasks performed by photojournalists right now, but as I noted last week there growing use might see a change in the very notion of journalism, a shift which might in turn better suit the algorithms and might put human journalists at a greater professional disadvantage. Realistically such a profound shift, if it occurs, is probably still a fair distance off, and honestly I don’t think the impact is going to be so negative, but I think it’s an interesting thought that this sort of technology might change the nature of our work even more profoundly than the advent of technologies like digital imaging and the internet.
Another possibility which is perhaps more likely is that we might be more likely to see is the rise of algorithmic journalism exacerbating existing trends which are already seen as undermining traditional photojournalism. Citizen journalism for example, and particularly citizen produced images and video of news events, might become far more appealing to mainstream news outlets if it were possible to set an algorithm to the task of sifting through the available images for the best examples, and even cross checking those images with other sources to confirm their veracity (an important task given the Instagram debacle highlighted here a few weeks ago). The ability to quickly mine high quality, accurate non-professional images might thereby further erode the appeal of, and market for, professional photojournalism.
The idea of algorithms checking sources sounds far-fetched but there are companies like Newzulu who already make it their business to offer similar services, only for now using people to do the hard work of verifying material before it is sold on to a news organisations. Given the right advances and the right impetus it isn’t so hard to see algorithms taking on this sort of role, and again the possible increases in speed, accuracy and efficiency would make them hard to resist. I’ve mentioned before the Visual Media Reasoning software package under the development by DARPA to perform similar functions in a military context, and such technologies have a way of eventually make it to the open market (including most obviously the technology that’s allowing you to read this post). Add the rather more achievable goals of an algorithm being able to automatically caption, edit and upload photographs, and it seems likely algorithms are going to take over at least some of the workload. Whether these advances when they come prove positive and free people up for more creative, interesting tasks (as proponents suggest they will) or whether they put people out of work (as detractors fear) will remain to be seen.
Thinking in the longer term we might be looking at a major shift in what we think of as journalism, and photojournalism. For a while now the trend in photojournalism has seemed to be away from simply producing good photographs, which after all is something that so many people can do now. Increasingly the thing that seems to set the best apart from the rest is the ability to think creatively and bring sophisticated and appropriate visual and conceptual strategies to their topics. Returning to the idea of ‘getting a job they can’t automate’, perhaps this really creative side of photojournalism will become increasingly important as the less creative arenas fall to algorithms. We’re already seeing this to some extent in written journalism). To make an even more extreme prediction, perhaps as photographs, video and the cameras that produce them become ever more ubiquitous, the need to send a photojournalist to location to document a subject might start to seem ever more archaic and cost ineffective.
Foggy as the future might be what’s been demonstrated time and time again is that what people can do, an algorithms can often do just as well. When you look at the way industries like finance have changed, from rooms of yelling brokers to a few server racks, the way is pretty clear. The brokers didn’t believe their jobs could be replaced by algorithms, because they thought they were a special type of professional. Creative, specialist, intuitive. They were also wrong. Anyone who thinks photojournalism is particularly different, should maybe start to think again.