Is artificial intelligence the future of content marketing?

Show menu

Artificial intelligence already automates a lot of marketing processes for us, but could it take over content creation? And should it?

Emily Tatti

As a content marketer in 2018, you’ve probably asked yourself if artificial intelligence will one day steal your job. The answer depends on a few things, including how sophisticated the technology becomes, and what ethical decisions we make in the future (just how adept do we want computers to become?).

While I can’t tell you if computers will achieve singularity and take over the world, I can speculate on how AI will influence content marketing, based on what the technology is capable of now, and where experts see it headed. The good news is, it seems unlikely AI will be taking over content production any time soon. Instead, it could enhance our work as marketers – if we are discerning about how we use it.

AI technology today

A futuristic robot

When it comes to marketing, artificial intelligence helps us perform time-consuming tasks like data analysis and email automation.

It might surprise you to learn that several high profile publications already use AI to generate content, too. Forbes uses a tool called Quill to write earning reports, and The Associated Press uses Wordsmith to write thousands of articles about finance and sport. These systems can’t produce articles autonomously just yet – in order to get something legible out of them, someone has to feed in statistics. Each system, which has learnt basic rules about grammar and structure, then determines what data has the most weight and repurposes it into an article.

For example, here’s something Quill wrote about a baseball game (via The Guardian): 

Twenty-seven Colonials came to the plate and the Virginia pitcher vanquished them all, pitching a perfect game. He struck out 10 batters while recording his momentous feat. 

Tom Gately came up short on the rubber for the Colonials, recording a loss. He went three innings, walked two, struck out one and allowed two runs. The Cavaliers went up for good in the fourth, scoring two runs on a fielder’s choice and a balk.

The limitations

As you can see, this content serves its purpose, but it isn’t perfect. It’s impressive that the software can add (pretty dramatic) flair like “vanquished them all” and “momentous feat”, but the piece is nothing more than a recap. Marketing copy has to spur the reader into action, and it’s clear AI isn’t capable of that yet.

It might be, in the future. Google engineer Ray Kurzweil predicts that computers will reach human level intelligence by 2029, which means we could hypothetically start automating our content production soon. As mentioned in our looking to 2018 blog, 79 per cent of Australian content marketers struggle to keep up with demands for content. On the surface, AI presents an exciting solution. But there’s a reason it sounds too good to be true.  

Audiences only act after consuming content they connect with. While data tells marketers a lot about what works and what doesn’t, we also have to rely on our intuition to pursue the content that most engages readers, whether it’s in print or online. We have to speak to experts, experience things firsthand and use emotional intelligence to give our writing nuance. Nothing will turn readers away from a brand faster than inauthentic content.

AI can’t do these things. It uses data, not logic, to make decisions, which leaves it alarmingly open to errors.

Take Tay, Microsoft’s Twitter chatbot, which was programmed to learn from its conversations with other users in 2016. Within 24 hours, it had turned from a polite pseudo-teenager into an aggressive Neo-Nazi because it was imitating the trolls sending it offensive messages – and Microsoft had to shut the whole thing down.

While AI will undoubtedly develop beyond this, Tay still proves how important it is to have human control over content. Currently, AI systems are capable of producing 2,000 articles per second. Imagine how much extra work would be involved in vetting and editing that.

How content marketers could use AI in the future

While I don’t think AI should ever manage our content creation, I do think it could support the process in exciting ways.

Producing content is time consuming because so much labour goes into it – it involves brainstorming, researching, writing, formatting, optimising, distributing and analysing. What if you could take some of those tedious jobs out? For example, several US-based companies like Frase and Knight Foundation are developing AI research assistants specifically for content marketers, to help automate research. These systems will have the ability to scan thousands of news sources in seconds, before presenting you with the most relevant information.

The possibilities for personalisation are equally exciting. Artificial intelligence is already used to send customers highly targeted emails (this is how Netflix suggests movies and TV shows). Imagine if your website could change its photos and articles according to the visitor’s demographics. Or if digital content automatically updated to accommodate a reader’s background knowledge (based on their Google search history). By using AI to marry content and personalisation, you could cater to your customers on a whole new level.   

It’s important to embrace innovations that can help make our jobs easier, and artificial intelligence could give us the time we crave to be more creative and strategic. However, while I doubt AI will be ready to write memorable content any time soon, we should nonetheless be wary of its potential to compromise the quality of our work. AI will be most effective if we use it to complement what we do as marketers, not replace it.

Emily Tatti, assistant editor