How to Detect AI-writing and Plagiarism

The internet hasn’t been the same since ChatGPT came out.

Now, you just can’t trust anything you read online anymore. 

November 30th, 2022, an AI tool so advanced that it could replicate human-like writing was released. Writers were terrified, readers were confused, and copywriters feared for their jobs. 

Since then, there has been a constant battle between authentic and AI-generated content.

Blackhat SEOs resurrected from whatever dark corner they were hiding in, and with a few prompts here and there, they could churn out any number of articles in just a matter of minutes. The world had changed, and we had to adapt.

Don’t get it wrong — AI isn’t inherently bad. 

In fact, it has revolutionised many industries and made our lives easier in ways we couldn’t have imagined. 

For example, some AI tools can be quite helpful when brainstorming, developing the basic structure of an article, or generating ideas. 

However, when it comes to full-fledged writing, AI just falls flat on its face.

In its attempt to mimic human writing, some of the AI-generated content may sound well-written and professional, but it lacks the authenticity and personal touch that only a human can provide.

Plus, there’s always the risk of plagiarism, getting penalised by search engines, and losing credibility as a writer or marketer.

So, how do you detect AI writing and plagiarism? 

Well, it’s not rocket science.

What’s AI Writing?

AI writing or AI-generated content is as it’s written on the tin — content written by an artificial intelligence tool, like chatGPT, Jasper, or CopyAI. These tools use machine learning algorithms to analyse existing content and generate new text that mimics the style, tone, and structure of human writing.

What’s AI Writing Detection?

AI writing detection refers to the process of identifying and flagging content that AI tools have generated. While it’s not always possible to detect AI-written content with 100% accuracy, robots do leave some tell-tale signs that give them away.

So, How Do You Tell if a Piece of Writing Was Done by AI?

Tools are one way, but not the only way. 

If your company works with an external team of writers, they may have unwittingly outsourced their work to an AI tool. 

But before you enlist the help of your trusty AI detector, we suggest you try looking for AI-ish clues yourself.

Some of the common signs of AI writing include:

• Repetitive sentence structure and phrases
• Inconsistent tone and style throughout the content
• Unnatural use of keywords or phrases for SEO purposes
• Lack of personalisation or relatable examples
• Generic, cookie-cutter content that lacks depth and originality

Repetitive Sentence Structure and Phrases

AI-generated content tends to be template-y.

For example, if you come across a piece of writing that always starts with “Are you tired of [insert common problem]?” and goes on to provide a list of solutions, chances are it was written by an AI tool. Humans tend to be more creative with their sentence structures. So, if it reads like a fill-in-the-blank template, then it’s probably AI-generated.

Inconsistent Tone and Style Throughout the Content

Humans can adapt their tone and style depending on the audience, purpose, and subject matter. AI tools, on the other hand, have a limited set of data and algorithms to work with, leading to inconsistencies in tone and style throughout a piece of writing. If you notice an abrupt shift from casual to formal or informative to persuasive, don’t ask for a second opinion — it’s definitely AI-generated.

Unnatural Use of Keywords or Phrases for SEO Purposes

SEO is important, but not at the cost of readability and flow. AI tools often overoptimise by stuffing keywords and phrases where they don’t belong, resulting in awkward and unnatural sentences. If you come across a piece of content that reads like it was written by a robot obsessed with SEO, chances are it actually was.

Lack of Personalization or Relatable Examples

AI-generated content lacks the human touch. It often fails to include personal anecdotes or relatable examples that can make the writing more engaging and memorable. Instead, it relies on generic information and data, resulting in cookie-cutter content that lacks depth and originality. 

It can’t capture emotions or paint a vivid picture like humans do. So, if you ever find yourself struggling to connect with a piece of content on a personal level, you know the culprit.

Inaccuracies and Inconsistencies in Information

AI lies a lot. Okay, maybe not intentionally, but they can’t always distinguish between facts and opinions or understand the context behind a statement.

They’re the know-it-alls of the digital world, but that doesn’t mean they always get it right. AI-generated content is prone to inaccuracies and inconsistencies, especially when dealing with complex or nuanced topics. You want to take everything it generates with a grain of salt and do your own fact-checking.
Lack of Creativity and Originality

AI’s biggest weakness is a lack of creativity. It’s great at mimicking and replicating existing content, but it struggles with coming up with original ideas and concepts. While the internal was already flooded with repetitive and unoriginal content, it’s safe to say that AI is making it worse.

And What’s Plagiarism

Plagiarism is an older cousin of AI writing.

When lazy writers use other people’s work without giving credit or pretending it’s their own, that’s plagiarism.

Imagine pouring your heart and soul into a blog post or article, doing all the research and everything, only to find out that someone else has copied it word for word and posted it under their name. 

AI content also falls into this category. Since a machine creates it, it technically can’t credit sources or acknowledge inspiration.

The Top 5 Best Plagiarism Checkers For AI-Generated Content

It’s time to unleash the fury of technology on…technology. Where our human brains struggle, these plagiarism checkers excel. Here are the top 12 best plagiarism checkers for AI-generated content:

#1. Originality AI

Writers hate originality with a passion. You’ve probably heard how it’s ruined many writers’ careers, forced others to rewrite everything from scratch, and even caused some to give up on writing forever.

Developed by a team of GPT-3 experts and content marketing specialists, Originality AI is specifically designed to detect AI-generated content and plagiarism. It uses advanced algorithms and machine learning to analyse text for originality and analyse it for AI-ish patterns.

This is where it gets worse for content millers (and better for genuine writers).

This AI writing detector not only detects AI writing and plagiarism but also sniffs our paraphrased writing of the same. So, even if someone tries to spin your content and make it seem “original,” Originality AI will still catch them in the act. 

That explains why many ‘writers’ hate. If content mill sites like Iwriter, Hirewriters … you name them, were to use this, they’d probably go bankrupt overnight.

Pricing

Originality doesn’t have a free plan.

Their base subscription plan starts at $14.95/month, giving you 2,000 credits.

One credit allows you to scan up to 100 words. 

So, if you do the math, that’s about 20,000 words at $14.95/month.

Alternatively, you can opt for their pay-as-you-go plan, which starts at $30 for 3000 credits (non-expiry). 

That’s about 30,000 words at $30.

#2. Copyleaks

According to a study conducted by Cornwell University, Copyleaks was validated as the most accurate AI detector on the market, with 99.1% accuracy with a model coverage that included Bard and GPT-4.

You can read the full report here: https://arxiv.org/pdf/2307.07411.pdf

Founded in 2015, Copyleaks has become a go-to tool for educators, professionals, and businesses looking to detect plagiarism and AI-generated content. 

Pricing

The basic (free) plan allows you to do basic scanning for chatGPT, Bard, and GPT-4-generated content without signing up. 

However, its extended features, such as batch scanning, multiple languages, and API integration, require a subscription starting at $9.99/month.

Overall, Copyleaks is a highly accurate and reliable tool for detecting AI writing and plagiarism in various forms of content. And with its affordable pricing plans, it’s definitely worth considering if bots bother you as much.

#3. GPTZero

Developed with the same technology used by OpenAI’s GPT-3, GPTZero is a powerful AI-writing detection tool specialising in identifying content generated by the GPT family of models. 

With its state-of-the-art algorithms and comprehensive model coverage, GPTZero bills itself as the gold standard of AI detection, trained to detect GPT4, ChatGPT, LLAMA, Bard, and other advanced AI models. 

Plus, its sleek user interface and user-friendly features make it a hit among users who value both accuracy and ease of use.

But what makes GPTZero truly stand out is its ability to detect AI writing and classify it as per the level of human intervention.

Pricing 

The classic version is free, offering unlimited scanning for GPT-generated content.

#4. GPTRadar

GPTRadar is another AI tool built based on the GPT-3 model. It aims to detect and classify AI-generated content, including GPT4, ChatGPT, OpenAI-DALL-E, and other advanced AI models. 

With its impressive accuracy and ability to detect content in multiple languages, GPTRadar is a reliable choice for identifying AI writing in any form.

The tool analyses your text and performs calculations on metrics like burstiness and perplexity to detect AI-generated content.

Pricing

The tool is available for free testing but with limited features. For full functionality, users have to purchase credit on a pay-as-you-go basis. 

The tool costs $0.02 per credit, each allowing for one scan of 125 words. 

That may seem steep, but can you really put a price on catching those pesky AI-generated texts?

#5. AI Detector Pro

AI Detector Pro is a popular choice among marketers thanks to its advanced capabilities in detecting and classifying AI-generated content.

It uses deep learning algorithms to analyse linguistic patterns and detect any traces of AI writing in text. 

But what sets AI Detector Pro apart is its ability to provide detailed reports on the level of human intervention in the AI-generated content, giving you a better understanding of its authenticity.

Pricing

AI Detector Pro has three plans available: Free, Basic, and Unlimited.

The Free plan only provides three reports.

Their basic plan goes for $27.98/month (now discounted to $13.99/month), with 100 reports per month.

Their unlimited plan is priced at $49.98/month (now discounted to 234.99/month) and allows for unlimited reports.

So, if you’re serious about detecting AI writing and plagiarism, investing in one of these plans may be worth it.