If you want to make a marketer squirm, sweat or just feel a bit uneasy at the moment, all you need is two letters.
AI. It’s sent much of the marketing community into a tailspin, frantically trying to get a handle on whether it spells new business opps and upsells, or apocalyptic doom.
While Artificial Intelligence (to use its Christian name) isn’t new, only relatively recently has it breached the barricades of our happy little world in content marketing. Now it’s here, and making itself right at home, what should we do with it?
Start your watch. Because in five minutes’ time, you’ll leave here comfortably clued-up on the uses and limits of AI, and specific tools like ChatGPT and Midjourney, in content marketing.
What is AI?
Let’s start by making sense of some of the terms and acronyms flying around, many of which – you’ll be relieved to hear – we don’t need to worry about for our purposes.
The umbrella term that’s relevant to content marketing is generative AI, which applies any time the technology is used to create content – text, images, audio or video. Natural Language Processing (NLP) is one of the methods AI uses to do that – algorithms that sift through text for the human-like characteristics it can skilfully, often uncannily replicate. Within these NLP models sit Large Language Models (LLMs) – where we find the (in)famous ChatGPT – that perform this processing on a massive scale, practised on vast quantities of material and with general capabilities that can respond to near enough any query you send their way.
That’s ChatGPT in a nutshell. Much more than the glorified search engine that some of the more cynical among us have suggested it to be. But there are two versions of the tool to know about. ChatGPT Free is nice to say and runs off the batch of data inputted when it was created in 2022. ChatGPT Plus doesn’t roll off the tongue as well, but its access to the live internet makes it more accurate and evolving, fed with the latest, up-to-date information.
ChatGPT is merely the poster child for a growing club of generative AI tools. If you’re looking for something to assist with visuals and creative content, there’s Midjourney from OpenAI (parent of ChatGPT) and Adobe Firefly, which can conjure imagery via the same prompt system. But AI is also increasingly an add-on or feature within established software – even good old Google Docs now has an inbuilt writing assistant.
How much can we trust AI?
AI has been met in some quarters with the kind of panic that greets alien invasion. Unnecessarily so, so far. We’ve already allayed fears of redundancy and career-extinction, but there are still some good-to-knows and words of caution.
It’s attracted most of its bad press for inaccuracy. Some studies have shown a declining accuracy rate over time, and reports have illustrated examples of ChatGPT succumbing to the same hallucination issues that can hamper any AI. Mainly, what these instances – many of which, it should be said, have been contrived to expose flaws rather than occurring naturally in everyday use – highlight is the need for careful wording of prompts and an awareness that it’s a fallible machine like anything else.
Then there are debates around copyright. Can a piece of computer-generated content be protected by the same copyright laws? A federal judge in America set the precedent that it can't. It’s all free game. Adobe Firefly combats related issues of plagiarism and unlicensed use of authored material by only drawing from its own stock library.
What can AI do in content marketing?
It’s for these reasons that AI is best used to support your own creative process, rather than generating final content to be sent straight out into the world.
Whether getting the ball rolling, or delving deeper into your own ideas, AI is a great source of inspiration. Given a robust, thoughtful prompt, it can suggest themes, points and angles, or content formats that might resonate with your target demographic.
Many marketers are using AI to offload the sort of time-consuming research tasks a computer can perform much more efficiently. It can provide high-volume keywords for your content or analyse trends for insights into your audience. Ask it for your target persona’s common challenges and needs, and you can map out a fuller picture of their pain points and more closely personalise your comms.
It lacks the inhibitions and self-consciousness that can paralyse a human creative, allowing it to produce rich springboards for your team’s own ideas. With so many visual aids to choose from, creative AI tools can craft heavily stylised concepts layered with details you might not have been able to create independently against time and budget pressures.
What can’t AI do in content marketing?
While AI might be capable of abstract, broad-vision imagery, those skills don’t transfer all that well to language. It’s currently not so valuable a tool for campaign concepts and ideation, throwing out lines that – to the human eye – are either far too vague or far too literal. It’s best put to use when working within what already exists.
Some creative marketing teams also lost faith in the software when the initial iterations botched human features, Frankensteining unearthly horrors with seven-fingered hands or legs rendered at strange angles. It’s improved since – it’s a fast learner – but is yet to nail focused, compact, concentrated creative such as logos.
The trick is to be as precise and detailed with your prompts as possible. This is trial-and-error tech, literally designed to learn, grow and improve the more it’s used. You’ve gotta hand it to it: it’s pretty impressive. And even if – even. if. – it really does end up putting us out of a job, it can generate a list of alternative careers in no time. I bagsy dog-walker.
Is AI the greatest concern in marketing right now? Something else keeping you awake at night? We asked marketers across the country, 'What's the matter?' Here's what they had to say.