When you hear about the automation of certain industries, you might think of assembly lines and robot arms, or Uber’s plan to employ a fleet of driverless cars. Rarely, however, do we think that such a thing could ever happen to industries that require a creative input, like the process of writing marketing content.
I thought the same, until I found myself whispering whaaaaat at my computer screen as I read this article about how a novel co-written by an artificial intelligence robot was able to win the first round of a literary contest in Japan.
No one is safe!
The novel was also co-written by a human, but he only supplied starting details such as the plot progression and character details. From there, the A.I. program crafted an entire novel essentially by itself.
Developments like these are less surprising when you consider how A.I. technology has been seeping into our culture for a long time now. When your phone recognises that a friend has asked you a question, for example, and prompts you with an auto-complete response of ‘yes’ or ‘no,’ we might not even notice that artificial intelligence is casually at work in our hands.
The truth is that the capacity of A.I. programs has made some astounding progress in recent years. The Turing Test, for example, can be passed with ease by the robots of 2017. The test was created by computer scientist Alan Turing in 1950, and required that an unknowing person talk to an artificial intelligence robot over a text-chat. If the person finished the conversation unaware that they’d just spoken to a robot and not another human, the A.I. was said to have passed the test.
An important critique of this test, however, was that it didn’t really measure the intelligence of a robot - only the capacity of a robot to imitate human processes. But as far as the commercial use of these programs goes, does the difference really matter?
More companies than you might think are already employing A.I. software to write articles, reports, and other online content. MIT Technology Review reports, for example, that the Narrative Science software, Quill, is churning out over a million words per day and is in use by clients such as Forbes, T. Rowe Price, Credit Suisse and USAA.
Narrative Science began as an experiment to see if computers could turn baseball scores into narrative reports, and from there, additional research has allowed the software to produce ever more mature and natural-sounding ‘stories’ based on inputted data. Nowadays the program has found incredible commercial use processing reams of financial data and producing sophisticated reports in seconds - something it could take a human day or even weeks to do.
Another similar software, Heliograf, has been in use by The Washington Post for some time now to supplement their journalists’ output. It was used to create over 300 stories for the Rio Olympics, for example, and could even post to their social media pages with score updates.
The uses of such technology continue to branch out - Uberflip analyses the content a user has browsed to recommend relevant articles. Seventh Sense analyses when clients are most likely to look at an email and automates your send-out process to ensure the best chance of the content being engaged with. And so on.
Much of the ‘storytelling’ ability of content-writing software is still in development, with a lot of the articles still sounding a little… well, robotic. They appear to be best for straight-forward journalistic reports and summarising data findings. But with the speed at which things are developing, it surely won’t be long before NLG (Natural Language Generators) become sophisticated enough to produce even more compelling content.
So what does this mean for the future of content marketing?
As already mentioned, A.I. programs can process information and churn out thousands of words of content in mere seconds, saving marketers on a huge amount of the research, writing and scheduling legwork it takes to keep uploading fresh content. This could free up time for marketers to focus on other initiatives, projects and client relationships.
Additionally, the profitability of such software can be impressive. The Washington Post’s Heliograf’s technology is sold through Arc Publishing starting at $10,000 a month with increasing packages up to over $150,000 a month. Such prices can be justified with the huge amount of content it can produce at high speed, however, and The Wall Street Journal reports that the CIO has indicated margins of 60-80%.
In short, A.I. technology has the potential ability to seriously diversify marketing operations, and those who are at the forefront are likely to hang on to a real competitive edge when it comes to efficiency and customer optimisation. That said, for now it appears to largely be an analytical and perhaps a ‘quantity over quality’ tool. An excellent way for marketers to supplement their output, or to take away some of the legwork, but not a complete replacement for human capability.
I’m reminded of a scene from iRobot where Will Smith’s character asks an Android how robots could be considered anywhere near as valuable as humans as, after all, ‘Can a robot write a symphony?’
The android tilts its head and asks him, ‘Can you?’
It’s a powerful moment and one that the internet particularly enjoyed, too…
Whilst you can’t equate most content to a masterpiece symphony, it should still require a degree of inventiveness, rather than simply imitation of what has come before. Perhaps when the day does come that A.I. programs can replace human content completely, it’ll be a sign that we seriously need to up our creativity in our approach to marketing.