Can AI Content Generation Replace Human Writing?

Tags

AI, Artificial Intelligence, content marketing

Back in 2015, I wrote a hilarious post about artificially created romance novels, whence the gem below. So, how are things 6 years on?

Photo: Gizmodo

Artificial intelligence (AI) has made automation the norm in more fields than ever. In marketing, AI has transformed the way we communicate. The use of machine learning in profiling audience segments and running programmatic advertising campaigns is an example of how AI transforms the way we can engage the audience at different times.

AI and machine learning are also trying to tackle another big hurdle in marketing: content creation. As content—and content marketing more specifically—becomes a key instrument in digital marketing, a golden era for content creation is dawning. However, might AI content generation replace human writing, thus depriving writers of much-anticipated jobs?

An Advancing Technology

Let’s start by acknowledging how advanced AI has gotten this past decade alone. Machine learning can now mimic the human learning process quite closely with unassisted learning. After defining specific patterns, AI can consume large data pools to learn at an advanced level.

This scaled learning process also allows AI to pick up new patterns and define better, more efficient ways of replicating those patterns. When machine learning is used in Natural Language Processing (NLP), the result is rather spectacular.

We all remember when Google demoed Google Duplex, the company’s NLP entity. Google Duplex was able to make a reservation over the phone, acting as a human assistant, and completely fooling the recipient of the call.

For text and article writing, the advances are more significant. Not only can AI construct complex sentences, but they can also now take context, personal cues, and the articles your audience engaged with the most in the past into account.

This result is that AI content generators are becoming more effective in creating content that attracts the audience, especially when you target a specific audience segment, and you have a lot of learning data for that specific segment.

More start-ups and enterprises are investing in this type of AI. Media companies like The Washington Post developing their own machine learning ecosystem for the purpose of creating a capable AI-based article or content writer.

AI Becoming More Accessible

The use of AI to generate articles and other content is not limited to large companies. Anyone, including small businesses and home users, can turn to AI-powered tools for help in writing good content.

Article Forge is one such popular tool. It started life as a content rotator and rewriting suite and developed into an AI-powered platform for content creation. Article Force uses deep learning, another subset of AI, to build an AI instance capable of writing a unique article.

That last part is very important. With minimal human input, Article Forge can construct an article about most subjects. This is not generic content either because the generated articles are readable and tailored to a specific audience segment.

Article Forge also accumulates its learning data. This means that the tool will get even better as more people—digital marketers and business owners—turn to the platform for help with writing good articles. It is only a matter of time before we see more breakthroughs.

Another popular tool is Zyro. Zyro is relatively new, but the tool has been developed over the course of several years. Like Article Force, Zyro relies on accumulative learning data and vast data warehouses to build a capable AI-powered content writer.

Zyro’s basic tool is available for free. I fed the second paragraph of this article into Zyro’s free content writing tool, and this is what it generated in return:

The idea behind artificial intelligent systems goes back at least to 2005 when IBM wrote its own version that would create an employee out-of-work list based solely upon job titles—think about how ubiquitous all these applications are now! The possibilities of using AI in marketing is endless, including in generating content for marketing.

Minor errors aside, that’s a pretty good paragraph, don’t you think? With some tweaking, AI can scale the entire content operations, allowing marketers to redirect resources towards other, more important things. Things like understanding the audience better, investing in better targeting, and reviewing marketing campaigns better.

It Has Its Limitations

However, AI won’t be replacing human writers any time soon. Yes, at its present rate of growth, AI will be able to write entirely personalized, highly readable articles in just a few seconds. The infrastructure capable of advanced machine learning and deep learning cycles are also becoming more available; they are becoming more affordable too.

However, one limitation is tone. It is difficult for machine learning to pick up on emotional cues. Even with the best AI, you still get that sterile feel when reading the article. That’s not necessarily a bad thing. A more scientific or sterile approach is good for technical articles. It just doesn’t work with personal pieces or op-eds.

Human writers have something that AI can never replicate: emotion. When articles are written by human writers, they are warmer, more relatable, and more pleasant to read. Readers also benefit from the life experience of the writers; the personal touches that they can add to the articles they produce.

“At scale, AI is a lot faster. However, human writers have a lot to offer too, including a more personalized approach in writing or depicting a subject,” explains Simon, head of content at Keen. And he adds:

The best path forward is for AI to augment human writers, so writers can be scaled up.

Simon also pointed out that AI is more than capable of automating other parts of content production. Content rewriting can be outsourced to AI, with a human editor putting the final touches. Variations of an article or a press release can be generated, resulting in more variations, thus scaling up the reach of marketing efforts significantly. Don’t be surprised to see agencies incorporating AI into other workflows, especially since search engines penalize duplicate content—content repeated verbatim across multiple channels.

That brings us back to our original question: can AI content generation replace human writing? For now, the answer to this question is a big NO. A tool, yes. A replacement, no. AI will continue to get better, and they’ll be able to write better articles, but AI will be more of a writing partner for human writers rather than a replacement for talent.

Happy writing!