There's a new trend in content creation, and it's going to change the world of content marketing as we know it. That trend is AI content generation.
The precursor to modern AI, machine learning, has been in use in various forms of content marketing for years now. Things like autocorrect and autocomplete are very basic forms of it, tools like Grammarly and other grammar checks use machine learning to refine their understanding of grammar, and tools like Clearscope give AI-recommended tips for improving SEO scores in content.
None of this took content to the level of being automatically generated for you, though. Up until recently, you would need source content and an article spinner to do that. Article spinners essentially just replace words with synonyms and are very obvious and easily detected by Google.
Now, language-model AIs have hit the stage. Jasper, ChatGPT, Sydney, Bard, and who knows what else is coming down the pipe all offer to generate content for you. Google quietly removed their rules against AI content generation (and now simply rejects automatically-generated content with no human oversight), so, of course, many businesses are looking for ways to leverage these new tools.
I'm not here to make a judgment about whether or not AI is good or bad. There are pros and cons to the tool, just like any other tool. It's likely to change the face of content creation, and there will be growing pains. For now, we all just need to navigate the situation as best we can.While natural language AI is an impressive tool on the surface, it has some serious drawbacks. Two, in particular, come to mind.
The first is that all of the major AI models are, well, single models.ChatGPT is trained on a massive batch of written works of a wide range of content and style, with a complex process of reinforcement and refinement developed to build it into what it is.
Anyone who uses ChatGPT is getting ChatGPT. Same with Jasper, Bard, or any other AI system.
What this means is that if you and someone else in your niche both ask the AI to generate a piece of content for you, you're going to get results that are very similar. They'll be "unique" enough to pass Copyscape checks, but they're still likely detectable by AI checkers that are being developed (especially those being developed behind closed doors at Google). They also will be distressingly similar to one another, such that your blog won't stand out from other blogs generating similar content.
The second issue is that AI can generate very good-sounding, good-looking content that is extremely wrong.Don't think of this as a user issue, a budget issue, or a design issue. It's an issue fundamental to the way current language models are developed. They prioritize conversational styles and cromulent sentence structure over deeper and more sensible content.
For a great example of this, look no further than Google's debut of their Bard AI. This is an extremely powerful example of why, exactly, you can't rely on an AI generator to create your content for you.
What happened? Google released the first test of their AI publicly in a commercial to demonstrate its ability. In the commercial, it was asked a series of questions, and it provided cromulent, good-sounding answers.
Those answers were also, in some cases, completely incorrect. The resulting errors cost Alphabet a billion dollars in stock price by demonstrating how unreliable AI can be.
All AI systems currently available have this problem. They need fact-checking from an expert. Imagine if your flagship blog posts made fundamental errors on easily-verifiable facts. How would that reflect on your brand?
The good news is, you can hire freelancers to help you out with these and other AI-related issues. In general, there will be four roles you can hire for to utilize AI content generation to its fullest. Which one you choose, or if you choose multiple, depends on your workflow and how you want to use AI in your content marketing.
The first option is to hire someone, likely a freelance writer, to help you with using AI tools to create content.
This is a role that a lot of freelance writers are going to turn to in the coming years. AI can do the drudgework of writing low- and mid-quality content, filling out boring stuff like product descriptions and form letters, and other tedious, formulaic, or basic content.
For example, have you ever wanted to put together a glossary of industry terms as a piece of content for your site? Filling out one of those is tedious, time-consuming, and boring for a talented writer. An AI can generate that content for you, with some guidance from someone who knows two things: what terms should be on the list and how to get an AI to generate definitions for those terms.
Over time, this role may also expand into professional tool usage. Jasper, the biggest of the business-focused content AIs currently available, has its own specific workflows used to generate certain kinds of content in chunks using a process to guide the AI more definitively than a free-form "just ask a question" style conversational AI.
So, if you need someone to wrangle an AI to generate the kinds of content you want to use, and you need it to be done relatively quickly and fluently, hiring someone to use those tools for you is a good idea. After all, you may be fluent with your language and know your subject well, but if you don't know the conventions of writing, you won't be able to write your blog posts effectively; similarly, you may be able to use an AI, but unless you know how the AI system works, you're still going to be putting a lot of time into it.
The second role you'll need to consider is a content reviewer and editor. AI-generated content is impressive on the surface, but it can commonly use the same words and phrases over and over. It's also prone to over-explaining, repeating itself, and taking on a format that works stand-alone but doesn't work as part of a larger whole.
ChatGPT is a good example of this. Ask it a top-level question, and it will give you an overview answer. Ask it to expound upon one of the points it made, and it will compose a new reply that starts with an intro and ends with a conclusion, like a blog post in miniature. That's fine if it was going to be a stand-alone piece, but it won't be.
One of the larger drawbacks of current content AIs is how little they can generate at a time. Often, they will create 400-600 words of content at most. Since the average length of a good blog post today is 1,500-2,500 words, you need to use the AI to generate numerous individual sections and stitch them together.
A content reviewer and editor will take the AI-generated content and massage it into something usable. Removing duplicate passages, stitching together chunks of content in a way that is both logical and fluid, and interlacing elements of SEO, keywords, and other aspects of marketing into the content.
After all, the AI isn't going to include your calls to action, your tagline, or even your keywords by default if you don't tell it to do so. Knowing how to stitch all of these chunks of content together and make something usable out of them is a skill that will need to be developed, so hiring someone to do it for you saves you from having to learn it all yourself.
As mentioned above, one of the biggest problems with current AI is that it's much more concerned with readable, conversational results than it is with factual accuracy. That is to say, AI is wrong, very often. More importantly, if you're not an expert on the subject in the first place, you might not have the knowledge necessary to point out when it's wrong.
Hiring a fact-checker has been a beneficial role for some industries (like health and legal) for years, but now it will be more important than ever. This person will need to be a subject-matter expert, who knows as much as you do about your industry, and who can point out lines, paragraphs, and even conclusions that the AI makes and remove them if they're entirely wrong.
This is only going to get more and more prominent as AIs start to get trained on larger accumulations of content, some of which will have been generated by less sophisticated AIs and published without review. There will definitely be scandals and other issues as brands are caught publishing AI content with no oversight. Eventually, this will lead to tangible harm. Don't let your brand be the one to cause that harm.
Hiring a fact-checker doesn't need to be an AI-specialist role. Fact-checkers can review any claims you make and any sources you use, both in human-written content and in AI-generated content, and provide you with reports on what is and isn't correct. This can be a huge boon, especially in content where accuracy is required.
This one is a specialist role, but it may be more important for some industries. What is it? It's someone who has access to the tools and checks necessary to determine whether or not a piece of content was generated by an AI.
Some people will not want to publish AI-generated content. If you hire freelance writers, you'll want to know if your writers are creating content or if they're using an AI to do it for them. If you're accepting guest posts and contributions from others, you want to know if that content was generated by an AI.
This is going to be a very important role moving forward, and it already is in some cases. In recent news, the SFF magazine Clarkesworld – one of the most accessible magazines for new authors – had to close their submissions because of a flood of AI content. It's a crisis, and one that will make identifying AI content a very important task.
Unfortunately, this is a tricky role to fill because there are very few telltale signs of an AI writing content, and the currently-available detection tools are lagging behind the AI development itself. Still, it's a role that some of you will want to hire ASAP.
Some ambitious brands see that AIs like ChatGPT are based on an open-source neural network framework, and you may think, "what if we did that for ourselves? What if we had our own in-house AI?"
This is a valid question, and the answer you're going to run into is that it's perfectly viable to do if you have a lot of processing power and a lot of time or a lot of money to pay data centers to provide those for you. AI is not a simple, run-on-your-home-PC kind of software, as much as the web clients make it seem like it.
Still, some brands are going to want to invest in AIs that aren't easily detected, that are more specialized to their industry, or that are more heavily trained with factual accuracy in mind. Hiring an AI developer as a freelancer might be a good option if you're looking to take up that role.
If you're interested in hiring AI-specialist freelancers, I know a great place you can visit. My job board exists for all things content writing, and that includes not just writers and editors but the people who do those jobs for AI content as well.
Feel free to post your needs on the job board or look for work from someone who has their needs posted already. It's an industry that is only just now growing, so there's plenty of room for expansion.
