Since ChatGPT exploded onto the scene in late 2022, it — along with other large language models (LLMs) — have surged in popularity. In fact, some projections estimate that by 2030, up to 99% of internet content will be at least partially created by AI. That’s a problem for marketers whose livelihoods revolve around publishing original content, because human-generated content is key to successful campaigns.
To combat this issue, AI detection tools have emerged to help determine content’s most likely origin. However, as the sophistication of AI increases, so does the quality of content, making AI content detection more difficult. Even the most cutting-edge AI content detection tools run the risk of producing false positives that mistakenly identify human-written content as AI-generated.
False positives from AI content checkers can derail marketers’ content strategies, limiting their ability to effectively reach and engage audiences. Think about it: If a marketer receives a false positive result on a piece of content they planned on publishing later that day, they will think they have to go back to the drawing board to secure a fresh piece of content written by an actual human, wasting both time and resources and pushing the publishing timeline. The latest Google core update has amplified this challenge by prioritizing “original” content in search results — i.e., not AI-generated — and penalizing content it flags as being written by AI by torpedoing its rankings in search results.
To help keep their content on the first page of search results, marketers must leverage advanced AI tools that accurately distinguish between human-crafted content and AI outputs while simultaneously adapting their strategies to align with Google’s evolving priorities. Creating genuinely valuable, original content will resonate with target customers and keep search result rankings high.
Clarifying Content’s Creator
Figuring out whether content was created by AI or a human is no simple task because of the unending diversity of human writing styles and the ever-increasing sophistication of AI outputs. However, several factors can play into making the assessment. Accurately determining content’s origin hinges on the analysis of its perplexity and burstiness:
Perplexity
Perplexity determinations are based on the content’s unpredictability, which is inherent in human writing, such as overcomplicating sentence structures or inserting a fragment. The more perplexity in a piece of writing, the more likely it was written by a human — AI content is typically more predictable.
Burstiness
Burstiness deals with the variability of human writing. A human’s writing is likely to have sentences of varying length and different levels of complexity. AI, on the other hand, produces much more uniform, no-nonsense writing with sentences of similar lengths.
Measuring these two factors can help marketers decide if a piece of content was most likely written by a person or LLM. But even with these quantifiable details, assessing whether something was written by AI or a human is not foolproof, so it’s important that safeguards are in place to further help prevent false positives to most accurately determine content’s origin.
Tips For Telling Where Content Came From
Marketers can better prevent false positives by:
1. Implementing A Human Touch
Human editors are irreplaceable in determining if something is written by AI or a person. People, unlike AI, have intuition and empathy, which can help in deciding the most likely origin of content.
2. Using A Tool That Aggregates Several Content Checkers
Content checkers produce a probability score of whether content was AI-generated. Comparing multiple scores can give a better idea of the likelihood of it having been created by AI.
3. Reviewing Tone & Flow
Well-written content will have a natural flow and consistent tone, but beware — AI is always improving at mimicking human language patterns. Sometimes human writing can be stilted and inconsistent, too!
Combine Human & Tech Intervention
As stated, determining content’s origin is a complicated process, but layering technology tools and human intuition can help prevent false positives by allowing marketers to make the most informed decision possible. This approach is critical to prevent accidentally publishing AI-generated writing that will be penalized because of Google’s recent core update prioritizing original content. If Google flags marketers’ content as AI-generated, their search result rankings will drop drastically, losing both clicks and customers.
Moving forward, the smartest marketers will implement content-checking procedures that intelligently combine AI-powered content checkers and very human “gut feelings” to most accurately determine how content was created.
Andrew Kirkcaly is the CEO and co-founder of Content Guardian, an AI content detection platform. He is an experienced and results-driven brand, digital and E-commerce marketer. with more than seven years of experience working at the strategic executive board level for an FTSE 100 business. Kirkcaly has been in the digital and e-commerce field for more than 12 years, with 10 of those at AO World. He is now focused on building a new media publisher startup in the gaming and technology sector, By Gamers For Gamers, and Content Guardian.
