There are only a few industries in which automation isn’t threatening some job roles. That’s a pretty scary thought, right? Well, don’t panic just yet.
“While automation will eliminate very few occupations entirely in the next decade, it will affect portions of almost all jobs to a greater or lesser degree, depending on the type of work they entail,” according to McKinsey Quarterly.
Roles that require empathy, like therapists and psychologists, as well as jobs that are highly reliant on social and negotiation skills, like managerial positions, are less threatened by automation, according to The Future of Employment: How Susceptible Are Jobs to Computerisation?
Those of us in roles that require creative thinking and original ideas — like content creation — are also deemed at less risk of having our jobs swiped from under our noses by something harder-working, “smarter,” and cheaper to maintain.
For now.
It’s pretty tough to envision a machine generating great content ideas, not to mention creating that content — content worth consuming. Or so you might think.
The reality is that machines are already writing content — and they’re pretty good at it.
In fact, Gartner predicts, “By 2018, 20% of all business content will be authored by machines.”
By 2018, 20% of all business #content will be authored by machines via @Gartner_inc. Share on XWhile that’s only a year away, don’t panic — “business content” isn’t quite the same as creative content used for marketing.
Natural language generation
Natural language generation (NLG) is the name given to artificial intelligence capable of producing logical, coherent text.
“Natural language generation is a software process that automatically turns data into human-friendly prose,” as Automated Insights explains.
It’s clever but, unlike a human, NLG can’t produce prose on its own. The format must be templated, and it needs access to a structured data set.
For example, to use NLG tool Wordsmith, you upload your data, write a template and presto — you have content!
Want to see what the results look like? Chances are you already have, although you probably didn’t notice.
Have you ever read Forbes’ earning reports? They are generated using Quill, another NLG platform. Here’s what the content looks like:
“The consensus estimate remains unchanged over the past month, but it has decreased from three months ago when it was 39 cents. For the fiscal year, analysts are expecting earnings of $1.68 per share. Revenue is projected to be 2% above the year-earlier total of $369.4 million at $378.4 million for the quarter. For the year, revenue is projected to roll in at $1.56 billion.”
What do you think?
Sure, it reads fine and it makes sense. If you didn’t know it was written by a machine, you probably wouldn’t notice anything was off. But it lacks something.
The writing has no discernible soul, and why should it? A machine doesn’t.
Machine-generated content has no discernible soul, and why should it, asks @SujanPatel. Share on XThen again, this is financial content we’re talking about. It doesn’t matter whether the writing has personality. It just needs to accurately report the facts. And for that, NLG is ideal.
Let’s see something else.
Below are the opening sentences to two sports pieces (courtesy of the New York Times). One is written by a human and the other by a machine.
“Things looked bleak for the Angels when they trailed by two runs in the ninth inning, but Los Angeles recovered thanks to a key single from Vladimir Guerrero to pull out a 7-6 victory over the Boston Red Sox at Fenway Park on Sunday.”
“The University of Michigan baseball team used a four-run fifth inning to salvage the final game in its three-game weekend series with Iowa, winning 7-5 on Saturday afternoon (April 24) at the Wilpon Baseball Complex, home of historic Ray Fisher Stadium.”
Can you guess the author for each?
If you couldn’t, you’re not alone. A similar experiment using multiple pieces of text like those above concluded that “readers are not able to discern automated content from content written by a human.” (For the record, the second one was written by a human.)
The study also asked participants to rate each piece of content on 12 characteristics. The results are telling:
Software- and journalist-authored content (also known as machine and human) score pretty equally on factors like coherence and accuracy — characteristics that can easily be learned by a machine (or I assume they can, based on my limited knowledge of programming).
The machine-written content came out on top (noticeably so) for the criteria of trustworthy and informative. That’s fine — those things are important, but they’re not what make content “great.”
The human-written content, however, soared ahead in two critical categories. It was rated significantly less boring and significantly more pleasant to read.
That makes sense.
A computer can’t read the content with a critical eye, and it can’t understand the vital complexities and nuances of language.
But will it be able to one day?
Why Automation Is the Future of Content Creation
Turing test
Every year for more than two decades, the artificial intelligence community has congregated for the Turing test — a trial designed to determine whether machines are able to think and talk like humans. It’s named after its creator, Alan Turing — you might know his name from the loosely biographical Oscar-winning film, The Imitation Game.
In 2014, a machine won the Turing Test — reportedly for the first time.
Now, when I initially heard about this, I have to admit I was concerned. If a machine can fool people into thinking they’re conversing with another person, surely it would be capable of creating content that can fool people, too, right? After all, a conversation is unpredictable. A machine that can keep its cover when questioned should produce content using data and a template easily.
Thankfully (for those of us who don’t want to see AI replacing manual content creation, at least) things aren’t quite as they seem. For many years, the Turing test has been regarded as the benchmark for AI intelligence. If a machine can pass the test, it’s deemed to possess at least average human intelligence.
And yet, in the wake of this pass, a number of computer scientists and tech investors questioned the result. Scott Aaronson, a computer scientist and former faculty member at MIT, challenged Eugene (the “winning” machine) to a conversation. Here’s a snippet of what happened:
Anyone with a half-decent grasp of the English language should be able to see that Eugene is far from human. If this is what’s deemed equal to average human intelligence, I think content creators can rest easy.
For now.
Ray Kurzweil, Google’s director of engineering, believes computers will be smarter than humans by 2029. Specifically, he says, they will “be able to understand what we say, learn from experience, make jokes, tell stories, and even flirt.”
Just to clarify, this guy knows his stuff. Not only is he helping to bring natural language understanding to Google, but he has correctly foreseen similar things. In 1990, he predicted that by 1998 a computer would defeat a world-class chess champion. It happened in 1997.
Of course, Kurzweil isn’t talking about content creation specifically, but surely a computer that can comprehend language and learn from experience could create content that stands up to that written by human hands, right?
I honestly think it could — provided it’s given the right data. Based on Kurzweil’s predictions and the quality of content AI already produces, I have little doubt that in the not-so-distant future, computers will be capable of creating some pretty awesome content that’s indistinguishable from human-written content.
What I don’t believe computers will be able to do — at least, not our lifetime — is to think creatively. And that’s key.
Even if computers can create content, they will never be able to think creatively says @SujanPatel. Share on XThe point may come where machines are writing the bulk of business content and news reports, but could a machine write a moving opinion piece or a novel?
AI content creation is, for now, algorithmic. Its capabilities are based on the information we humans provide. This is where I think its limitations lie.
To fully replace manual content creation, AI has to be able to think like a human. It has to be able to feel (to have emotions), it needs to form opinions, and it needs to think critically.
Should that ever happen, I think we’ll have much bigger things to worry about than the demise of manual content creation.
What do you think? Do you believe AI will ever replace truly creative content creators? Let me know in the comments.
Please note: All tools included in our blog posts are suggested by authors, not the CMI editorial team. No one post can provide all relevant tools in the space. Feel free to include additional tools in the comments (from your company or ones that you have used).
Want to explore further the role machines can play in improving your content marketing today and in the future? Don’t miss the Intelligent Content Conference March 28-30 in Las Vegas. Register today and use code BLOG100 to save $100.