The robots are coming! At least to the journalism industry - specifically earnings and crime reporting.
Software systems, such as Wordsmith, are writing news reports about corporate financial results at speed and in volume. The AP estimates it is now covering the results of 4,400 companies with Wordsmith. Previously, it was tracking 300 human authors.
In fact, Wordsmith churned out 300 million stories in 2014 - more than every article written by humans in every country combined. That’s a lot of content.
As a fan of automation in PR, I was asked recently what I thought about this trend. Are we all about to be replaced by robots, as we’ve seen in manufacturing? In short, no. Here’s why:
Robots aren’t creative. While there will be improvements in the tone, pace, style, and voice of automated content, at present it’s rather limited and dry. Feed the algorithms enough big data examples of what we consider to be first-class writing and it’ll be able to produce similar patterns of language and writing which could convince you it was created by a real person. However, it will still be a replica - not the original. Robots won’t have the insight, understanding, humor, or humanity that characterizes good content.
It won’t be topical. Brands only win the war for relevance by being topical. Don’t expect an algorithm to comprehend the zeitgeist of the moment. That’s why we’re still seeing Volkswagen banner ads next to articles about Dieselgate.
It won’t be visual. Google’s Jetpac uses image recognition to automatically produce city guides, but the opposite isn’t possible. Automated content systems don’t take an idea and produce a visual which captures that concept succinctly. The trend in communications is towards more visual content.
To be clear, I am in the camp of those that support robot journalism. There is value in creating articles which otherwise would not have been written. At present, those are information-dense, time-specific pieces that were largely repetitive, error-prone, and had low value for reporters to write themselves. Given the changes in the media landscape, I doubt it’s a major concern for journalists if this type of reporting goes away. In fact, it frees them up to focus on the why and how rather than the what, when, and where.
Equally in PR, while I don’t think it’s directly applicable today, ‘robot PR’ is likely only to take a supporting role at the bottom-end of content creation. I can’t imagine, for instance, that there is a lot of value drafting virus reports for security companies. Those might fit the time-specific, formulaic style appropriate to a system like Wordsmith. When it comes to PR, there’s a bigger opportunity for automation in monitoring and analysis than content creation.
Looking at other areas of marketing, automation software has focused on optimizing distribution of content and analysis of customer behavior rather than creating the content itself. You might get some keyword suggestions, but the system isn’t drafting the blog posts or email campaigns just yet.
So, if the robots are coming to PR, they’re not going to steal your pen. At least not yet.
Then again, shift happens.
PS: However dry, dense, or formulaic you might find this post, I can assure it was written by a real person. No doubt a robot could have done it in half the time.
Morgan McLintic is an EVP at Lewis PR.