At first glance, copywriting seems easy to automate.
Sometimes it feels like most blog posts, marketing reports and outreach emails were written by a computer, anyway. Surely they are just copied from one another, with a few changes to avoid charges of plagiarism, and then published, right?
Well, yes and no. It may be true that a lot (and perhaps the majority) of marketing content is generic—but not the best content. For that, you need a human touch.
1. Auto-generated content is on the rise.
Fears that artificial intelligence (AI) will replace copywriters have been around for a while. Back in 2017, the Oxford University Future of Humanity Institute released a report called ”When Will AI Exceed Human Performance?” in which they asked 352 AI experts their opinions.
The results make frightening reading for copywriters. Within 10 years, the report asserted, AI would be able to produce passable high school essays, and within 25 years they might be producing convincing novels. Presumably, somewhere within this timeframe, AI will be tasked with producing marketing content, and it will do this much more quickly—and much better—than we poor humans.
Well, maybe. At the moment, at least, text produced by AI is pretty easy to spot, and pretty hard to read. “Article spinners”—or automated tools that take existing articles and change the wording so they can be reused—have been in use for a few years already. They are still not very effective, and they often require a human copy editor.
2. Will AI writers replace human writers?
This is not to say that AI isn’t already having a huge impact on the marketing sector. AI in marketing is already here, and if you work in the industry you should prepare for the onslaught of AI and emerging tech.
Most high-quality marketing software, including latest-generation DIY site-building programs such as Wix, already incorporates AI-driven tools to improve customer targeting and engagement. JP Morgan Chase made news when it signed a five-year deal with AI content firm Persado. Yet when it comes to producing high-quality content, marketers will be relying on humans for some time to come.
To see why, it’s worth looking at the newest AI writing tools available. In Japan, for instance, an AI has been used to produce novels that, in one instance, passed the first screening phase for a major literary prize. The Washington Post has also experimented with these technologies: In 2016 the paper used a tool called Heliograph to compile reports on what was happening in the Olympics. It’s rumored that Reuters is also investing in AI in order to write simple news reports.
Basic newswriting tools are developing quickly but still have a long way to go before they can replace human writers. One reason is that they all rely on significant levels of human intervention in order to work. Washington Post journalists were tasked with “driving” Heliograph during the 2016 Olympics, for instance, rather than just letting the AI do its work. It required constant support and maintenance.
There have also been high-profile failures that temper zeal over the use of AI-generated new content. The USGS, for instance, tried to use AI to generate earthquake reports, but turned it off again when it started warning residents in California about an earthquake that had happened in 1925. Given that political, military and business decisions are made in response to news reports, it would seem foolish to let AI write the newspaper.
Finally, there is another important reason why AI won’t produce killer writing anytime soon: It doesn’t know what it’s like to be a human. That might sound like needlessly philosophical, but it’s important. AI researchers have long realized that for AI components to develop true intelligence, they must be contained in a body and experience all the attendant constraints.
3. A cooperative bond lies ahead.
To connect with your customers, you must provide a relatable human voice.
Your customers might be happy to get weather reports compiled by AI, but they are not going to buy a car (or anything else) from one. For that, they must trust the person selling it to them, and as humans we trust humans more than machines.
That’s not to say that AI can’t help us when it comes to producing and marketing written content. AI is already used extensively to assess the efficacy of written content and other marketing materials, for instance. Still, this is always done at the request of humans who know what they are looking for. Rather than having AI replace human writers, then, it’s more reasonable to assume they will work together.
There are already many examples. AI is revolutionizing many aspects of the digital world, including cybersecurity and the way data centers work, but in both cases AI components are under close supervision. Chatbots are another great example of this co-operative relationship. Chatbots and AI can transform your internal comms, and most customers are happy to talk to an AI in order to get answers to simple questions. Still, customers want to be able to talk to a human about more complex queries, or if their inquiry is of a sensitive nature.
4. AI will be (almost) everywhere.
Ultimately, the reason AI won’t replace human writers is the same reason they won’t replace human nurses. When it comes to big decisions—about our health or about our purchases—there’s no substitute for the human touch. No matter how convincing a machine is at imitating human behavior, humans won’t trust that a machine has their best interests at heart.
Dan Fries is technical product lead at Next Ventures.