Is AI killing copywriting?

Jun 7, 2019 by
Is AI killing copywriting?

Whilst the ‘rise of the machines’ heralded by every science fiction film from The Terminator to The Matrix might still be a few years away, many industries are already starting to feel the impact of creative artificial intelligence.

Every industry under the sun is going out of its way to embrace AI, but though this disruption might be celebrated in the short-term, in the long-term it could lead to a not-so-distant future where it’s that much more cost-effective for businesses to use to AI to replace human workers.

Most traditional ‘blue-collar’ roles are safe (for now), but in the creative industries, those wanting to create high-quality content at scale might soon start turning to AI over tried and tested designers, online marketers and copywriters.

This is, of course, a scaremongering, worst case scenario that might not ring 100% true, but there are legitimate concerns to be addressed. Because AI might not be killing copywriting, but it is certainly changing it.

Made in China

The writing was on the wall (please pardon the pun) last year when Chinese e-commerce giant Alibaba developed an AI that could potentially create 20,000 lines of content per second.

Granted there was no way to tell if this content was useable, but for businesses wishing to create product descriptions and similarly ‘dry’ content, it could mark the start of a serious sea change for the copywriting industry.

The tools have already supposedly passed the Turing test and companies like the fashion chain Espirit are reportedly already using the tool, which uses “deep learning and natural language processing technologies learn from millions of top-quality existing samples to generate copy for products.”

Make no mistake, this certainly poses a threat to copywriters whose bread and butter lies in bulk, inbound marketing work, but we are still many years away from being able to produce an AI that can create content with genuine character.

Even as far back as 2015, agencies were experimenting with this tech, with M&C Saatchi designing a poster ad that essentially rewrote itself in order to engage passers-by.

According to chief innovation officer David Cox, the idea was for the poster to “write itself based on what works, rather than just what a person thinks may work.” Alibaba, however, is the first company to really utilise the widespread marketing potential of the tech.

What is NLG?

Natural Language Generation uses AI to identify the focus and the general intent of a piece of content and then rewrite that content in a new way that mimics the tone of the original. Companies are essentially teaching AI’s to write like humans, and it’s working.

NLG tech might have originally been developed to produce analytical reports, but in recent years, the potential for using it to create short copy has been discovered and there’s no turning back now. Let’s be honest – the vast majority of short-form copy is often rather generic, which means it’s that much easier for an algorithm to work out what makes it tick.

AI is also uniquely able to work out which specific phrases generate the most engagement and create content optimised around those phrases,  potentially in seconds. It’s also a technology that has obvious beneficial implications for headline and keyword generation. For longer content, however, we might be safe. For now.

Robotic Replacements

It’s likely that AI will continue to play a much greater role in how we produce copy in the coming years, but whilst NLG can be utilised to create content with clear and simple intent, it has yet to fully comprehend the complexities and the ambiguities of human language.

So, human copywriters might have a competitive advantage for the foreseeable future, as AI is relegated to doing the ‘grunt work’ of the industry – predictable content.

Is AI killing copywriting? No. Or if it is, it’s doing it very slowly indeed. However, the day when an AI can craft a compelling narrative to rival the greatest writers of all time is perhaps closer than you might think.

Tags: