Narrative Science raises $6m to replace web copywriters with computers. How long until journalism is automated?

By

The internet can be characterized very simply. Billions of people are looking for interesting and useful information, and millions of companies are trying to make money by people finding their content, through search engines and increasingly on social media.

This has led to the rise of companies such as Demand Media, which last week listed on New York Stock Exchange to be valued at $1.5 billion, more than the New York Times. Demand Media and its peers such as Associated Content, now owned by Yahoo!, use low-cost writers and sophisticated algorithms to create massive amounts of content tailored to generate revenue from search traffic.

There are also many writing brokers such as TextBroker and The Content Authority that help smaller companies that need web content to improve their search rankings to get copy written, at rates as low as 1.2 cents a word.

I have written about the proliferation of crap content and how search is evolving to deal with the rise of low-quality content. The latest iteration in Google’s search algorithms explicitly address duplicate content. The quest for original content to feed the search engines continues.

The obvious next step is to automate copywriting, further improving the cost-revenue equation for those seeking to attract search traffic.

Enter Narrative Science, which has raised $6 million to do what it calls on its website “Quality content created automatically, at scale.2”. It already has eneratedgStatsMonkey, which writes stories about baseball games, and supports sports news site StatSheet, which I wrote about in the rise of robot journalists. The next step is to write stories about anything, or rather things that are in demand and easy for computers to write, such as real estate listings, financial reports, or hot new websites.

This is clearly just a first step. It will be interesting to see whether search engines will be able to distinguish the better computer-written text from human-written text. It’s hard to see that they could, given that the lowest human-written text is so poor.

Beyond this computer writing will get better. For search engines, the issue is not whether content is written by people or machines – it’s whether it is good quality content, in other words useful to searchers. Increasingly computer-written articles will do that.

As I always note when I talk about the future of work, there are two things to consider in any job:
* can it or will it be able to be done remotely, by people anywhere in the world?
* can it or will it be able to be automated, by computers or machines?

Some types of writing are falling into the second category. That doesn’t mean humans will be supplanted – it means that we will need to write differently, pushing us into realms that distinguish our humanity.

As Peter Kafka of MediaMemo puts it, writing about Narrative Science:

The trick for content-makers like myself is to find work that only content-makers like myself can do – work where human qualities like experience, judgment and creativity get rewarded. And if we can’t do that, we ought to be doing something else, anyway.

These fall into the ‘Novelty’ and ‘Insight’ categories described in my NewsScape describing value creation in the new media world. So, to answer the question in the headline, some types of journalism are on the verge of being automated today. However the really worthwhile writing and journalism will probably never be replaced by computers.