Can a computer program put writers out of work?
By Laura Haight
Originally published as The Digital Maven in Upstate Business Journal, Greenville, SC
“Tuesday was a great day for W. Roberts, as the junior pitcher threw a perfect game to carry Virginia to a 2-0 victory over George Washington at Davenport Field.”
If you are a regular reader of Forbes or a fan of Big Ten sports, there’s a good chance that you’ve been reading stories created by computers.
In fact, you might be shocked at the amount of content that is being “generated” by computer programs — and who they are writing it for. How about the Associated Press? The world’s largest news service. They’ve been using Automated Insights to write financial reports. The AI Wordsmith system is also marketed as a way to produce client results reporting for marketing companies.
The Big Ten Network has been using another company — Narrative Science — since 2010 to do in-game updates and tweets for basketball and football.
And its not just publishing services. It’s possible that the annual client report you get from your financial advisor has been produced by an algorithm under the umbrella function of a Natural Language Generator (NLG).
Think you could tell the difference? Try this quiz from the New York Times. I won’t even tell you how many I got wrong.
This really isn’t new. It’s been going on for more than five years and, as you may be starting to see, is gaining acceptance in some major industries and some big players. Some in the NLG industry predict that 90 percent of all content could be created by robot programs by 2020.
Narrative Science says 2015 is the year of the “democratization of data” where collecting information gives way to deep analysis and utilization — performed, of course, by artificial intelligence.
The NLG movement segues nicely with the rush to transform “education” into “skills-based training” and the promotion of technology over those-soft skills of critical and analytical thinking.
The confluence of internet adaptation, cloud computing, content marketing and a recession have created a huge contingent force of freelancers writing blogs, books, white papers and articles for publication. Will a computer program be able to replace them?
Already high quality writing and reporting is devalued for cheap and quick. Good writers struggle to compete with $5 blog posts available through freelance bidding sites online. Of course, you can really tell when a blog post has been created by someone who wrote it for $5. It lacks context, uniqueness, reporting, flow, tone, and so much more. But many cost-resistant small businesses want blog posts only as SEO magnets for high search ratings and referrals.
I would argue that whatever the reason you have content on your website, that content still represents you. If it’s slapdash and shallow, that’s what you represent to the viewers you have managed to attract to it.
We are awash with content (Five ways to get more hits, 15 things you need to know about hiring, the three things you wish you could take back) that is the written equivalent of bubble gum. Sweet, not filling but ultimately bad for your teeth.
The algorithm-generated content that I have read while preparing this article (yes, it is really me) is better than a lot of the content I see posted on LinkedIn, Facebook and Twitter. But still stays in its lane: reporting on data points. At least for now, to go beyond that, you still need an educated human.
