Artificial intelligence is among the hottest trends in today’s business environment. From handling customer service to predicting consumer behavior, it has simplified many tasks and unlocked a wealth of opportunities.
However, machines and algorithms aren’t just supporting professionals in many industries but replacing them altogether. According to the Brookings Institution, 36 million jobs have been disrupted by AI in the US alone, as above 70% of the tasks performed in them can be fully automated.
Until a few years ago, AI was mostly taking over repetitive and mundane tasks, freeing up room for creative work. But now, machines seem to have lightbulb moments of their own. The emergence of machine learning (ML) and the concepts stemming from it are allowing algorithms to develop a new skill – writing.
Yes, AI writers are now a thing, and they’re getting better as you’re reading this. So how far can the technology go? Will human writers join quills and typewriters as relics of the past?
To answer these questions, we must first look at the technologies that gave machines their first creative sparks.
NLP and NLG – The Driving Forces of AI Content Creation
Listing out all the branches of AI would take a while. But when it comes to writing, there are two main concepts that you should know about.
The first one is Natural Language Processing (NLP). In essence, it allows computers to understand written or spoken language and respond to it.
NLP is widely used in everyday gadgets. If you’ve caught yourself using phrases like “Hey Siri” or “Hey Google,” you’ve already encountered it.
Over time, NLP has evolved far beyond virtual assistants. With new advancements, the technology is getting better at understanding not just individual phrases but also the context behind them.
So how does it tie into content automation?
Think of NLP as an AI writer’s eyes and ears. It receives input and structures what it received in a logical manner. On its own, it’s not a human writer’s competitor. But combine it with Natural Language Generation (NLG), and you get a full-fledged writer.
As the name implies, NLG allows algorithms to transform structured data into a narrative, either written or spoken. It forms a symbiotic relationship with NLP to write content similar to that of a human.
The question is – how similar?
AI-Generated Content in Action
Over the past few years, machines have gotten impressively good at generating content. At this moment, there are a few capable solutions out there, mostly used by large newspapers and media companies.
Let’s have a closer look at some well-developed AI writers and what they can do.
Heliograf is a news bot developed and used by The Washington Post. It got its start in 2016, initially used to cover the Rio Olympics. Since then, Heliograf has generated hundreds of small reports.
While it’s not exceptionally creative, Heliograf does what it’s supposed to do. It provides sports updates, automated earnings coverage, and similar content. So far, Heliograf has proven extremely accurate, which is why some of its human counterparts consider it a threat.
Source: The Washington Post
However, The Washington Post claims Heliograf isn’t there to replace writers but to help them eliminate repetitive tasks and take care of work done in bulk. While it may be true, the same was said of a range of AI-powered bots that ended up taking over jobs like customer care. We’re yet to see whether Heliograf can evolve beyond short reports and perfect its writing skills. But judging by some of its more advanced colleagues, this is far from impossible.
If you need proof, just look at another AI content creation tool that shows how far the technology has come.
Narrative Science is easily one of the strongest players in the content automation space, and Quill is its flagship product. It’s a platform that turns raw data into cohesive and digestible narratives.
Here’s an example of what Quill can do:
“The energy sector was the main contributor to relative performance, led by stock selection in energy equipment and services companies. In terms of individual contributors, a position in energy equipment and services company Oceaneering International was the largest contributor to returns. Stock selection also contributed to relative results in the health care sector. Positioning in health care equipment and supplies industry helped most.”
Of course, Quill relies on human input, but the narrative that it can generate is quite impressive. This is especially true when you take into account that Quill’s strongest suit is writing at scale. According to Kris Hammond, founder and CTO of Narrative Science, Quill is able to mimic the style of a sportswriter and generate 1000 times more articles than a human!
Quill has written millions of words so far, and it’s constantly evolving. Companies and even some federal agencies are already using it, so the platform surely contributes to the idea of AI writing content on its own.
Not impressed by reports and news updates written by AI?
What about novels?
Booksby.ai is a platform that sells books made entirely by AI. If that’s not futuristic enough, the bookstore itself is created by AI as well!
The entire library looks like something straight out of an episode of Black Mirror. According to the founders, even book reviews and images of the reviewers are generated using algorithms without a single element designed directly by humans.
As for the books, they were created by exposing AI to massive numbers of science fiction books. Through training, algorithms were able to mimic their tone, style, and even visual appearance to create an entirely new book.
A perfect example is Dinner Depression. Both the contents of the book and the cover were designed by AI. In fact, the book even set its own price for which it’s sold on Amazon.
Now, do these books make for a smooth read? Not really. None of them is perfectly comprehensible, so reading them is mostly a matter of figuring out what’s going on. But when you take into account how recent this technology is, it’s almost scary to imagine how far it could go.
This is especially true since machine learning makes algorithms increasingly independent. Instead of people having to perfect book-writing AI, the algorithms can improve themselves through experience.
So not only can AI write formulaic content, but it can also get somewhat creative. The result might not be what you expect, but AI takes the phrase ‘practice makes perfect’ to a whole new level.
What does all this mean for writers? Is it time for content agencies to take the “human” out of “human resources” and start comparing features instead of resumes?
Writing Is Still Very Much Human
Even though AI seems like a recent trend, its beginnings reach further into the past than you might think. The concept of an artificial neuron surfaced back in 1943. Since then, scientists and tech experts have had one goal – to emulate the human mind.
When you look at AI right now, there’s no doubt that we’ve made astonishing progress. And yet, we were never able to digitize certain traits that make us human, traits that are omnipresent in writing.
So let’s see some of the main reasons why machines can’t replace content writers.
1. AI Can’t Tap Into People’s Desires
Content marketing serves many purposes, with sales among the key ones. Today, just about everyone in the business world knows that people buy with emotion.
Sure, you might rationalize your decision. But pick any random thing you own, and you’ll see that there was an emotion that made you buy it.
Emotions are what makes us alive, and machines can neither comprehend nor evoke them, at least not on their own. It takes a great deal of human input to create an illusion of emotion in a machine. This is why an AI writer can’t compete with a human when it comes to content geared towards selling.
Take product descriptions as an example. There are several AI-powered platforms that can create them with minimal human effort. But not terribly convincingly. At best, you can expect a formulaic description structured in a cohesive narrative.
Here’s an example of a description that could be written by AI:
“The Rolex Day-Date 118206 is made out of high-quality platinum. The size of its bezel is 36mm, and it features an ice blue dial with Roman numerals. Other prominent features include the Day-Date Mechanism and President bracelet.”
Compelling? Not really.
Now compare the above to the following description:
“The Rolex Day-Date 118206 embodies power and prestige. The striking luminosity of its platinum case is perfectly accompanied by a gentle ice blue dial, which infuses a little softness into the piece while still radiating class.
The platinum hour markers shine from their elegant background, adding even more sophistication to the watch. The legendary President bracelet is paired with a concealed Crownclasp that gives the lavish timepiece a little subtlety.”
If you were a Rolex enthusiast, which of these descriptions would do a better job of persuading you to buy? The latter, of course.
Because you don’t buy a Rolex for its features; you do it to indulge in its opulence, celebrate your success, or simply send a message of power.
A machine can’t understand this, as it doesn’t know what it feels like to wear a Rolex. It can define the President bracelet, but it can’t appreciate its rich history and the value that the wearer sees in it.
Simply put, AI can’t connect to the very source of purchase decisions.
Of course, this doesn’t only apply to luxury products. It takes a meaningful human connection to understand your audience, their needs, and how those needs change over time.
Despite its outstanding progress, AI is nowhere close to matching people’s ability to do this. The best it can do is offer data that will help you figure it out. So from this perspective, we’re not talking about competition but collaboration.
2. People Relate to People
Since AI can’t understand human desires, it can’t relate to them. And relatable content is the backbone of successful marketing.
If you need proof, look no further than influencer marketing. In 2019, the industry was worth around $8 billion. According to projections, its value could jump to $15 billion by 2022.
So why is influencer marketing so popular?
You see, influencers aren’t just experts in their field. They’re individuals whom people know, like, and trust. The most successful influencers have a personality that people can relate to and a lifestyle that the audience vicariously lives through. And the content they create reflects this.
You can see the value of relatability far beyond influencer marketing. Business coaches and mentors tell rags-to-riches stories to make themselves relatable and give people hope that they can achieve incredible success. Brands humanize themselves through values and opinions to avoid coming across as faceless institutions. All of that involves tapping into the core of our humanity.
An algorithm can’t love, hate, agree, or disagree. It can only try to emulate all of this through extensive (and expensive) human input. In most cases, the end result would still require some tweaking to get that “oomph” that creates genuine connections.
So, while you can automate content creation to an extent, you can’t virtualize a personality that someone will truly relate to. That gives humans lots of points on the scoreboard.
3. Imagination and Creativity Are (Still) Reserved for Humans
AI content tools offer the same benefits as any other evolving technology – more work in less time. Alibaba’s AI writer can push out a whopping 20,000 lines of copy per second. Forbes’ Bertie can compile error-free reports much faster than any person.
While this is undoubtedly impressive, content that compels isn’t about quantity but quality. And the quality of content that you create depends largely on creativity.
AI still relies on existing ideas and historical data. It can’t deviate from what it learns or break the rules. In other words, it doesn’t have the imagination necessary to create engaging content.
Books written by AI are a huge leap forward in the quest to ignite creativity in algorithms. But they’re still art for the sake of art. You’ll see proof of this if you read any of the books available on Booksby.ai, or even just the summaries. While algorithms can display a form of creativity, they can’t arrange and direct it in a way that conveys a clear message or engages the reader.
High-quality content that meets a specific goal happens at the crossroads of imagination and structure. And this is something exclusive to humans.
A Human Touch Is Irreplaceable
Since the dawn of time, stories have been the cornerstone of our evolution. Written or spoken, words that traveled through generations resulted in the world we live in today.
But those words aren’t mere pieces of information that you can throw into an algorithm. It’s the emotions and meaning we attach to them that move us forward. In other words, it’s our humanity that gives context to the content.
For a machine to take over writing in full capacity, it would need traits that are fundamentally human. Creativity stems from imagination, relatability from self-awareness, and connection from empathy.
AI developing these qualities is a controversial outcome. On one hand, it’s where we want the technology to go. On the other, it’s exactly what we fear, as it would enable the old cliché of robots taking over the world.
But let’s not go that far into doomsday scenarios, but rather answer the question you’re here for.
No, AI can’t replace human writers, at least not for a very long time to come. Like many other practices, it can only enhance writing if used well. If we befriend modern solutions instead of competing with them, the synergy can give our words more power than ever before.