Fight AI: Shout out to the MC

Intelligent clients are increasingly checking their content for AI. Are you? If not, you need to add an MC to your QC.

Imagine the situation: You’ve paid a writer to create a new blog post (great!). They’ve delivered the work, and it’s a solid 1,500 words, reads well, and meets the brief. Let’s get this blog on the site, right?

Wait a second… At 42group, alongside our QC checks, we’ve added a new stage – our machine check, or MC. Every piece of content gets MC’ed before we recommend it’s ready to publish. The rise of AI writing assistants means we’re checking *everything* before we share with clients to ensure it’s 100% original.

If you’re commissioning content, here’s why you need to add an MC check to your content production process (or work with an agency that does!).

Content at scale

At 42group, we create content at scale for our customers. For example, we recently produced 50 pieces of SEO-optimised content for an AI start-up in less than four weeks. (Lots of coffee, late nights, and attractive bonuses.)

The easiest thing do was to use an AI copywriting programme to take some of the strain out of creating briefs, structures, lists, and more. But that would have been the worst decision for us, our client, and their investors – and we’ll explain why.

SEO agencies and cut-price content agencies are quick to promote AI because it’s a (relatively) simple and cheap solution. In a market as competitive and cutthroat as search, why wouldn’t you?

Because SEO is all about Google – and however sophisticated you might think your AI writer is, its algorithms are just as smart. And they don’t need to be that smart to pick up on the sort of cliché-ridden junk produced by most AI writing platforms.

Despite what the LinkedIn evangelists say, AI is a pretty predictable language model that’s easy to unpick.

Assessing the damage of AI

Everyone is asking one thing: Will AI content impact my Google ranking?

Let’s see what Google says.

“Using automation—including AI—to generate content with the primary purpose of manipulating ranking in search results is a violation of our spam policies.”

The key words here are “manipulating ranking” – that (to us and Google) means publishing page after page of junk content. Or rather, content with no value.

Google uses advanced AI checkers to spot this type of content. If you’re using ChatGPT or Bard (or any other tool based on these platforms) to create content without editing it, you could be hit with a ranking penalty.

Some writers (and agencies) have suggested that AI can help with article structures, research, and even be used for writing parts of articles (such as lists of examples). Some agencies will tell you that by simply rearranging the structure of articles or briefly editing content, you can trick Google.

But you can’t.

You are what you E-E-A-T

While it’s true that Google hasn’t explicitly said it will penalise AI-generated content (at least for now), even stuff written by AI that you subsequently edit won’t add value. If it doesn’t add value or provide a unique perspective, then it could harm your site because it doesn’t deliver on any of Google’s four content ranking factors.

Google uses the acronym E-E-A-T to describe its ranking factors. If you don’t already know, it stands for:

  • Expertise
  • Experience
  • Authoritativeness
  • Trustworthiness

If you’re using AI to create content, you’re not demonstrating expertise because you’re using previously published source material. You’re not providing insights from your experience but simply crowdsourcing it from others.

AI creates content with no authority. It’s often written in a passive way, with no authority. It adds nothing to the conversation; it weakly amplifies the echoes of other posts.

There is a growing trend for stories about AI mistakes. We’ve seen sources misattributed and incorrect links. We (and Google) can’t trust anything AI says or tells us. Couple that with the fact that most AI platforms use a historic data set that’s over a year old, and it’s no wonder the results are often irrelevant.

On every level, AI content fails to meet the E-E-A-T test.

Using AI isn’t about adding value to your readers and your site. It’s the opposite – and it could damage your website, business, and brand. (So don’t do it!)

MC like a pro

We’ve built a reputation for delivering work that’s shaped around audiences and algorithms. We call it building human connections, and it’s a rallying cry for the industry.

It’s not a meaningless slogan; it’s a principle: only human writers can kind of content that moves real people.

We have a team of writers, some in the office and others who work remotely. As part of all projects, we QC everything. And, since the start of the year, we MC all work, too.

Our Machine Check process has three stages that analyses:

  1. The originality of content (whether it was written partly or wholly by AI)
  2. The sources linked in all content
  3. Plagiarism

To help us, we use several tools, including OriginalityAI, Grammarly (and its plagiarism checker), and ChatGPT’s AI content checker.

Each piece of content is analysed in several ways to assess its originality and value.

We’re not going to run you through the process (as it’s boring, naturally) but having one in place means we can provide all customers with the confidence that they’re getting what they paid for.

Is your content agency doing this? If not, they should

When commissioning content – particularly large-scale blog projects for SEO-purposes – make an MC a key part of your approval process.

If you want to chat about content, AI, MCs or just have a natter about old-school hip-hop, then get in touch.

Want to connect with your customers?

Get in touch!