How to Spot AI Content in the Chat GPT Revolution 

Tread Luddite-ly 

Imagine you are a Middle School Teacher who has decided to teach a unit on the Industrial Revolution. To evaluate your students’ mastery of the subject matter, you have assigned a 2-page paper, and are just now starting to review their submissions. At first glance, you are quite pleased to see that nearly all your students filled both pages with observations, definitions, and interpretations that demonstrate a comprehensive understanding of how the Luddites banded together to destroy the textile machines threatening their livelihoods. But then it hits you: a lot of these papers seem similar in their content, tone, and cadence. You can’t quite put your finger on it, but they read like old encyclopedia entries. And since nobody uses encyclopedias anymore, that can only mean that a substantial portion of your students likely utilized AI writing tools to complete their papers. Your first instinct is to run the papers through an AI detection tool, but then you remember that a popular AI content detector recently flagged the U.S. Constitution as AI-generated and is therefore unreliable. You don’t want to make any conspiratorial accusations of wrongdoing like some boorish Luddite, but you also don't want to enable academic shortcuts that will only hurt your students’ abilities to write and think critically for the rest of their lives. So, what do you do? 

 

Who Stake-holds the Stakeholders? 

Before we answer that, let’s switch perspectives. Now, imagine that instead of a Middle School Teacher, you are the Director of Marketing for a $50M+ B2B Software organization who is reviewing content submissions from stakeholders for marketing distribution. And instead of the Industrial Revolution, the content covers a wide range of topics relevant to your company’s target market. Just like the Middle School Teacher, you have a sneaking suspicion that a large portion of these submissions were created using AI, due to their uncanny wooden structure, tendency for repetition, and total lack of brand personality. Except that now you don’t have the luxury of comparing pieces that are covering the same subject matter. They're all different. And they’re coming from different departments within your organization, which means the format, structure, and knowledge base is highly variable. What’s worse, you know that AI content not only has the potential to hurt your brand’s search rankings - as major search engines have begun penalizing indexed pages with lots of AI-generated content - but it could also reflect poorly on your position as Director of Marketing. You have the final say on the quality and quantity of marketing materials your organization is bringing to market. Like the Middle School Teacher, you don’t want to have any uncomfortable conversations with coworkers who may have written these pieces from scratch (maybe they were STEM majors), but you also don’t want to lose your job in 6-12 months when the real impact of AI-generated content embedded in your marketing ecosystem is fully known. So, what do you do?  

 

I, For One, Welcome Our GPT Overlords 

We are living through historic times. AI-generated content is here to stay, and AI technology is already saving lives in hospitals, autonomous vehicles, and even human trafficking syndicates. However, this revolution in language, word, and thought processing does not come without its share of pitfalls, as we’ve seen in the two scenarios above. So, how can Marketing Directors, Educators, and everyone in between, reliably check whether the content they are consuming has been created by an algorithm or a human who possesses competency in each area of expertise? The easiest, and most direct, method is to ask the stakeholder questions about the content they've submitted. Not only will this help identify individuals who have no practical understanding of the topic and likely used AI to write their piece, but it will also help reinforce key concepts with those stakeholders who really did write their content from scratch. Questions like:  

  • What are the key takeaways of your essay/blog/whitepaper? 

 

  • What makes these takeaways unique? What are you saying here that hasn’t already been said? 

 

  • What do you think some of the key objections to your arguments might be, and how would you respond to them? 

 

  • When you set out to write this, what surprised you about the writing process? Was there an ‘ah-ha’ moment that happened between the outline phase and the final draft? 

 

  • What do you hope the reader will do or think after reading your piece? 

 

Spot the Bot 

Of course, you may find asking the above questions to be awkward and even confrontational. If you work in an organization with delicate inter-departmental politics, hitting a stakeholder with an AI interrogation may not exactly make you popular around the office (even if you work from home). So, short of asking your colleagues to show, not tell, their mastery, how can you spot AI-generated collateral with your own two eyes? Here are some of the tell-tale signs of AI-writing:  

Repetition  

Repeating ideas or concepts using different words or phrases throughout a paragraph is a dead giveaway of AI assistance. 

Lack of Analysis  

An encyclopedic statement of facts, in a cookie-cutter thesis statement and supporting argument structure, generally indicates the content was formulated by AI mining various sources available throughout the internet. 

No Clear Voice 

Anyone who has worked in a corporate environment long enough knows that stakeholders tend to use insider terms and jargon that may convey a unique brand personality. If these aren’t present, it was likely written by a bot. 

Non Sequiturs 

AKA hallucinations, non sequiturs are a byproduct of content written by an algorithm that is simply attempting to formulate the closest approximation to a piece that will be recognized as genuine by human beings.  

 

Conclusion 

AI content is here to stay. To avoid the negative consequences of confusing AI-generated words with critical thinking from human beings, ask questions, look for the telltale signs, and if all else fails, hire WB Creative Consulting for all your content and copywriting needs ;) 

Previous
Previous

The Power of Copywriting in Tech or: How to Convince Your Engineers that Marketing Isn’t Useless