Ay Yi Yi AI: Red Flags That What You’re Reading is AI-Generated
How likely is it that you’ve seen an AI-generated image or text recently? Pretty high, actually. And multiple sources say that by the year 2026 (that’s in two years, y’all), 90% of content on the internet will be generated by AI. So how can you trust what you’re seeing on your social media feed these days? Don’t worry. There are plenty of content marketing agencies (much larger than Next Page) that have been looking for ways to spot AI-generated content since… well, since AI became a topic in the mainstream conversation.
Here are a few red flags we suggest you look for (and share these with people you know and love! It’s an election year, after all).
🚩Red Flag #1: Content is Repetitive and Predictable
The reality is, if you stop to read the content you are seeing online you can probably tell which content is not written by a human. How? According to Semrush, you are looking for perplexity and burstiness. This means you are looking for sentence variety. Long sentences, short sentences, contractions, etc. If you are reading content and you can almost predict what the next word choice will be (often, adverbs are a dead giveaway), then it’s a good sign you are reading content written by AI. Adverbs are weak writing and AI generators depend on them because they are easy “filler” words. PS: Adverbs are words that end in -ly if you need a refresher.
For example:
We regularly have new products in the store: Tells you something basic.
versus
Be the first to know about new products in the store by signing up for our newsletter: Removes the adverb (regularly) and directs the reader to take action.
🚩Red Flag #2: Lack of Depth and Emotional Personality
People have plagiarized work since the beginning of time. I was most aware of this as a writing teacher in the 2010’s. It wasn’t that hard to find plagiarized work if you knew to be on the lookout for it. Because the people who are lazy enough to plagiarize are also too lazy to hide their plagiarism. As a teacher, I could, most of the time, Google a few key phrases from a student’s paper and sure enough the website would pop up with the exact phrases highlighted.
Now, it’s a little harder to tell if what you are seeing is AI-generated, namely because it’s not quite as obvious. It may sound like it was written by a person (albeit a boring person with too many adverbs). While there are tools like Undetectable and Sapling to help people determine which content is AI-generated, you may not always know to check those sources. This is where empathy and personality play an important role in content. When we onboard new businesses, we want to know what problems their customers are facing (among other things). This helps us write content that is empathetic and understanding to those hurdles.
Another example:
Call us if your employee clicks on a phishing email: Gives a simple directive after a problem happens.
versus
Employees may not know what to look for in a phishing email, and unknowingly open themselves and your business to cyberattacks. We’re here to provide education and guidance to stop phishing attacks before they happen. Empathetic to the problem and provides a proactive solution.
It’s easy to catch emotionally “unintelligent” writing if you start looking for it, especially when you read it next to more “human-centered” content.
🚩Red Flag #3: Outdated or Incorrect Information
Yeah, this one sucks if you are someone who doesn’t like to do research. Our team works to verify content before we place it in a blog or social media post, but not everyone has the time or patience to do that. And if you’re hoping that ChatGPT or another AI content generator will help you with this, you’re going to need a time machine. Because what’s “new” or “trending” is not something AI can help with. It is only as smart as the content it knows - and many of these platforms are only as smart as two years ago. So the issue here is two-fold: first, be sure to verify the content you are seeing by multiple sources. And second, make sure that any content you share or post (especially if it is connected to your business) is verifiable so others will trust you as well.
One final caveat to this red flag: your content must be also sensitive to your audience. When we first explored AI in 2023, we found that content which was generated for a very specific audience (in this instance: high-earning professional Black women) was generic and, to be honest, insensitive to this particular audience by not touching on the very real stigmas that this audience experienced. Not talking about that bias doesn’t make for a trusting relationship, because not talking about it suggests that you are either oblivious to it (not good) or emotionally unmoved like, say, a robot (also not good if you want people to trust you).
The bottom line: do your due diligence on the type of content you are seeing - even and especially if it is sent to you from a trusted family member or friend. If you can’t verify it, or it sounds just… off, then tell the friend/family member who sent it to you. Knowledge is power. And remember that social media is the perfect breeding ground for AI content to spread because of what the platform is built to do: encourage users to stay on the platform by consuming more and more content. #KeepScrolling is probably tattooed on Mark Zuckerberg’s arm (Just kidding. There is no verifiable proof of that).
Stay safe out there everyone! Until next time.