Will AI Be the Future of Content?

Published in

Writers Guild

5 min readNov 12, 2018

Imagine you are visiting a website from a software vendor and you want to contact customer support. You click the button at the bottom right of your screen and you write a question, hoping to get an answer in what you deem a usual time — anything from 2 to 24 hours. After all, a human needs to get to your message and write an appropriate answer to your problem.

To your surprise, after you press Enter, you get an immediate answer with a link to a page with a solution to your question. Boom! Problem solved.

Is that a fantasy? Not really. The current state of machine learning has developed bots able to understand what someone writes and provide a pre-determined response to it.

While there’s the possibility that a bot will not provide the kind of custom-made answer you’re looking for, the trend is real: machines can understand what we’re saying. If they can break down our speech and give us something, even if it’s a short answer, theoretically they could also write a piece of content.

I mean, we already have a machine capable of printing a new Rembrandt-like painting. Why couldn’t a computer analyze all my writing and come up with something that does not only resemble my style but can replace me?

What can AI do for writers?

Imagine you work at a financial institution, managing the portfolios of several wealthy clients. Every week, you are supposed to send them a summary of the performance of each client’s stocks. The problem starts with the fact each client has a different portfolio, and each stock has its own story to tell. One went up for one reason, the other went down for another one. Creating a clear report for each client would take you not less than 5 hours, and you manage several dozen clients. What would you do?

Hiring a full-time writer is the best option, but that can cost you a lot of money. What’s more, you can’t wing it by delivering second-hand reports because your clients are paying a hefty sum to have you manage their portfolios. They want the best, and you must deliver.

This represents the ideal opportunity for machines to shine. The area that specializes in generating AI-powered content is known as Natural Language Processing (NLP for short). With the help of technologies such as morphological segmentation and syntax analysis, a computer can take a set of structured data and create a coherent narrative that makes them look as if a human wrote it.

Take a look at the report Quill, a platform developed by Narrative Science, is able to generate for a wealth management client:

Good news about your retirement investments
Since last quarter, your retirement account has grown. Your balance increased 2.7% to $43,898 at the end of September. You’re also up 3.2% for the year.
You’re still on track for retirement
We’re watching your portfolio and by our calculations, you’re on track to reach your goal of $300,000 for retirement in 30 years. To assist you in staying on track, we have digital budgeting tools available online at URL
According to their website, Quill “automatically analyzes, interprets, and communicates information from your data the way your business does.” Their client roster includes Credit Suisse, USAAA, and different U.S. intelligence agencies. According to an MIT Technology Review article, Quill currently “churns out millions of words per day.”

Another company in this space is Automated Insights. Named one of the world’s Top 25 Artificial Intelligence Startups by CrunchBase, this company also generates AI-generated content, including creating content descriptions for fashion online retailers, hotel descriptions, real-time stock analysis, and profit-and-loss summaries.

Other companies in the space focus their efforts in other complementary tasks, like Frase, which calls itself a “AI-powered Research Assistant,” Acrolinx, which offers an AI platform that “uses a unique linguistic analytics engine [that] reads your content and guide you to make it better,” and Yseop, which offers a multi-lingual, enterprise-level language generation tool in English, French, German, Dutch, Spanish, and Japanese.

Will writers lose their jobs?

In 2016, McKinsey analyzed over 800 professions in the US and discovered machines could automate 45 percent of the activities, while 60 percent could see 30 percent or more of their jobs automated. Such a grim forecast doesn’t add a positive light on the role AI will play in the future, even for writers.

As we’ve seen, AI can write only from specific data sets. Feed an algorithm with data like a match score, weather information, and names of people, and you can get a wonderful match recap, like Automated Insights was able to generate.

But will a machine find the obscure assassination of Russian spies in the UK and come up with a Pulitzer-level investigation? Not likely. Humans are still too attached to the content creation process to imagine a computer write a piece of content on its own.

A machine may be great for a Wikipedia-like content where you need lots of facts. But for stories? No, thank you.

AI + writers = a match made in heaven

The current situation writers face is hard to explain. While the technologies that could replace writers already exist, they aren’t ready to make the leap — at least, not yet. But they could, and that possibility is what makes the potential grim outcome harder to imagine or expect.

So far, we can safely say AI can compliment writers — and that’s the great news. A machine can help a writer find facts, like stats, quotes, and other time-consuming tasks that could be automated.

But they won’t take our jobs. AI can write content and copy that is based on facts. What it cannot do is deal with human emotion in writing.

Before we build any walls against robots, let’s embrace them. Let’s use them for the reason they exist: to help us become better at our craft.