While machine learning and artificial intelligence (AI) algorithms have been in development and use for several decades, their impact became unignorable when consumer-facing AI products like Midjourney and ChatGPT became available in 2022. These products showcased the capabilities and applications of AI models to the broader public, and since then AI has been increasingly integrated into daily life for individuals and businesses. 

Discover 18 actional tips for offering white-label solutions to clients. Get the “How to white label” guide for FREE today!

As exciting as this new technology is, it also raises concerns about things like intellectual property and the authenticity of work. AI detector tools and services have emerged promising to analyze content and detect output produced by AI models rather than humans. This leaves AI users wondering how to avoid AI detectors, and whether AI checkers are accurate at all. In this article, we’ll take a look at where the evidence stands on detection tech for AI and whether it’s possible to bypass AI detection. If you use AI in your agency work, these are important considerations to be aware of.

How accurate are AI detectors?

AI detectors can be accurate, but they have serious limitations and can’t be relied upon to be 100% precise in their assessments. While a variety of tools have emerged making claims of high-accuracy AI detection—some even assuring users of a nearly 100% accuracy rate—many tools run into the issue of producing false positives. AI-generated content may accurately be identified, but human-generated content can still be interpreted as at least partially produced by AI. 

For example, a user tested several detection tools including industry leader WinstonAI, which claims a 99.98% accuracy rate. While the AI-generated text was correctly flagged, the human-generated text was identified as 75% human. The AI checker was directionally correct in that it identified the text produced by ChatGPT 3.5 as AI-generated and the human-written text as mostly authentic.

However, in situations when the stakes are high for identifying the authenticity of work, identifying human-produced text as 25% AI-generated can be a problem. That margin of error wouldn’t be sufficient for, say, an educator to confidently penalize a student for submitting AI-generated text rather than authentic work.

To use an example closer to home, if a client were to use AI detection to verify your agency’s work, they could easily turn up false positives. That’s why it’s critically important to be aware of how accurate AI detectors are—in a world that increasingly relies on this technology, it’s important to educate your clients on the possibilities, limitations, and potential inaccuracies of both AI and AI detectors.

How do AI checkers work?

AI checkers work—interestingly enough—in much the same way AI tools work. That is, they use algorithms, machine learning, and natural language processing (NLP) techniques to “learn” the patterns inherent to natural human text and AI-generated text. Just as AI tools like ChatGPT and Bard are trained on real-world information and human writing, AI detectors are trained on large datasets containing both types of information which help the models identify text that follows the patterns most commonly associated with human writing or AI-produced writing.

In the world of generative AI, two concepts repeatedly come up as tools for gauging the quality of output: perplexity and burstiness. These concepts also underlie how AI checkers work. If you want to learn how to avoid AI detectors, familiarizing yourself with these is a must.

Understanding perplexity

Perplexity is a measure of how accurately an AI can guess the next word that should appear in a sentence. This is developed with pattern recognition based on the analysis of large datasets, and the lower the perplexity of a model the more correct its output will be.

Consider an example: Given the sentence “I chewed some gum and blew a big (blank)”, an AI with low perplexity would likely guess the word “bubble”. Most of the time, when you hear this phrase in real life, that is the most likely word to follow. A less likely but still correct next phrase might be “wad of cash”. An AI that shows mid-range perplexity might come up with that phrase based on context, but it’s less likely. An AI with very high complexity might say “rollercoaster” — that doesn’t make sense, and it wouldn’t be a useful, practical AI model. 

As you can see from this example, NLP models exhibit a sweet spot in terms of perplexity. If they’re “overfitted” to the data on which they were trained, they will only work for narrow use cases but won’t generalize to other applications. A general AI tool like Chat GPT must exhibit enough perplexity to apply to different subjects, while an AI used internally at legal firms should more narrowly conform to its training data. 

Understanding burstiness

Burstiness is a measure of how regularly words, phrases, and syntax patterns appear in a text. Natural human text is generally “bursty”. When we speak, we vary the length of our sentences. We might follow a long, complex, run-on sentence with a short, simple one.

Text produced with AI models tends to be more consistently rhythmic and, well, robotic. Certain words or themes may be repeated more than in natural human text. Let’s take a look at an example.

A high-burstiness model might produce the following sentence:

“The unmistakable aroma of freshly baked bread wafted through the house, filling the air with warmth and anticipation. It smelled so good that I woke up from my deep sleep. I was salivating.”

A low-burstiness model might look more like this:

“I woke up to the pleasant, inviting aroma of baking bread. The baked bread was soft, warm, and comforting.”

The first example shows varied sentence lengths and a variety of words, while the two sentences in the second example have a similar length and structure. The phrases “baking bread” and “baked bread” exhibit a repetitive pattern that doesn’t feel particularly natural. 

These two measures—perplexity and burstiness—are used by AI detectors to score output. Generally, high-perplexity, high-burstiness text is more human, while low-perplexity, low-burstiness text is more likely to be AI-generated.

How valid are AI detectors?

The proliferation of AI has marketers, especially those working in SEO, asking themselves how accurate AI detectors are. 

Today, AI detectors face limitations in accurately discerning between AI-generated and human-produced content: while they can be useful for spotting AI-generated output due to familiar patterns from NLP models, they struggle when it comes to human-generated content, leading to false positives.

It’s easy to understand why this is the case. Biases and limitations of AI detectors arise from the fact that even human content, especially when written for the web, can be relatively low in perplexity and burstiness. We may speak in complex, unexpected bursts, but we tend to write in a more formulaic way. 

A 2023 study evaluating various AI content detection tools confirms this, showing higher accuracy in identifying AI content (especially when produced by GPT-3.5 rather than GPT-4), but lower accuracy for human-generated content. 

Should you bypass AI detection?

Even if AI detectors aren’t foolproof, if you have an AI startup, you should prioritize bypassing AI detection. These tools are likely to become more sophisticated, and while the role of AI in digital marketing will only grow, successful agencies will leverage this powerful new technology ethically and carefully.

The reality is that AI is here to stay, and there’s no need to constrain your business by eschewing it altogether. Different scenarios and use cases demand varying approaches to bypassing AI detection. Let’s take a look at a view.

• AI marketing integrations that respond to reviews: Since the authenticity of these isn’t a direct ranking factor for Google, agencies don’t need to expend extra resources humanizing them to bypass AI detection. Just focus on choosing a trainable tool or white label AI software that produces acceptably human-like responses.
• AI content marketing: It’s important to humanize AI content, especially when producing it for clients, because getting flagged by Google can severely impact their rankings.
• AI social media automation: AI has many practical applications for social media that don’t come with AI detection concerns, such as social listening and automated posting. When creating content for social media, it should be humanized 
• AI social marketing campaigns: Images and text produced for social marketing should be reviewed to limit AI detection.
• AI art: Ensure you use AI art generation tools that comply with copyright and IP regulations in your region.
• AI lead generation: Lead gen tools aren’t generally a concern when it comes to AI detection, so you can use them freely to grow your agency.

Want more insights on how to train AIs to better conform to your brand requirements? Check out this video:

 

How to bypass AI detection with genuine AI content 

Now that you understand how AI checkers work, the key principles behind how to avoid AI detectors should be more clear: 

• Content modification: Switch up the language, structure, or style of AI-generated content to make it less likely to be flagged.
• Customization of AI models: Use AI tools that can be trained or adjusted to your specified parameters to keep the output on-brand.
• Humanization: Don’t underestimate human creativity, which can be introduced through personal anecdotes or unique associations that are unlikely to be made by AIs.

It’s sometimes recommended to “trick” detectors by obfuscating AI content with intentional errors that appear like human mistakes, but this isn’t a good approach for AI marketing agencies since it can damage your reputation and make you appear unprofessional.

Conclusion

AI is here to stay, and so are AI detectors. Knowing how to avoid AI detectors ethically and reliably is an important skill for today’s digital agencies. Since this technology will continue to evolve in leaps and bounds, it’s a good idea to stay on top of new advancements both in AI and detection technologies, so you can make the most of these powerful new tools.