Hit enter to search

How to Detect AI Writing


Joanna Yuen

Marketing & Content Specialist - 19 Apr, 2024


Unless you’ve been in hibernation, you no doubt have heard the term “AI”. Artificial Intelligence, or AI, refers to computer systems that can perform tasks that typically require human intelligence. That means that many areas of problem-solving, learning, and decision-making can be handed off to machine learning rather than entrusted to humans. As you can imagine, that holds a lot of appeal.

One of the most popular uses of AI is for writing. In fact, over 85% of AI users say that they mainly use the technology for content creation and article writing. AI can help people who are not comfortable with their writing skills to level the playing field. From brushing up your resume to creating a killer LinkedIn profile, and writing ad copy to drafting an offer letter on your dream home, AI provides the tools that can help novice writers to make their work stand out.

But is there a downside to all of this AI content? There are actually several potential drawbacks, including a lack of creativity and human touch and an increased risk of plagiarism. This is especially true if you are paying for content creation in some form - you want something that is personally crafted, and not generated by a machine. How can you tell the difference? In this post, we are discussing what you need to know about AI, including how to spot AI generated content.

The Rise of AI

AI is becoming more popular for several reasons. First of all, advancements in technology have led to an increase in both computing power and storage capabilities. This has allowed for more complex AI models to be developed and deployed. On top of that, the availability of large amounts of data has enabled different AI systems to learn and improve their performance through a process called machine learning. Yet another reason for the growing use of AI is the extent to which these systems can automate and optimize different processes and tasks. This provides ample opportunity for increasing efficiency and productivity, while streamlining work tasks. What company doesn’t want projects to be completed faster, and what person wouldn’t want to do their job in the simplest way possible?

AI is being used in various industries from healthcare, finance, and manufacturing to transportation and logistics, just to name a few. Furthermore, AI has the potential to provide personalized experiences and recommendations to users. This is sure to become a powerful way to enhance customer satisfaction as well as engagement. These advancements are already being seen in applications like virtual assistants, recommendation systems, personalized advertising, and intelligent chatbots. Overall, savvy technologists will be looking to AI to tackle more complex problems, improve efficiencies, and provide more customized experiences, which today’s consumers have come to rely on.

Just how pervasive is AI? ChatGPT, the main technology that powers several machine learning programs, was launched in 2022. In just five days, it reached 1 million users. For reference, when Instagram launched, it took 2.5 months to reach a million users, and Spotify took 5 months to reach a million users. According to Reuters, ChatGPT set a record for the fastest growing user base.

In fact, AI is growing so quickly and so expansively, that some people are asking if perhaps we are moving too quickly. Instead of humans needing to advance AI to try and speed up technology, we are faced with adjusting other processes to keep up with AI. For example, AI is likely to be used by students to create term papers, and therefore teachers will need to find new tools that can identify AI content. While up to 85 million jobs will be eliminated by AI, up to 97 million new ones will be created within the next two years.

AI can be used to provide a competitive advantage, create better workflows, and improve customer experiences. It can also provide several small advantages which can help even small companies to squeak out an edge over their competition. Some of the most popular applications for AI are in the following tools:

  • ChatGPT - This is probably the one you’re most familiar with, as the innovative technology actually powers other tools. The original AI platform, ChatGPT from OpenAI represents about 60% of the AI tools in the market. Currently the most popular AI option in use,
  • Character.AI - This tool is currently the most-used chatbot. Dialog agents are built from large language models and are created with conversation in mind. You can create characters and have them span applications like entertainment, education, customer service, and more.
  • QuillBot - A paraphrasing and summarizing tool, this technology helps millions of students and professionals cut their writing time by more than half using state-of-the-art AI to rewrite any sentence, paragraph, or article. Users can simply enter a prompt and explore the suggestions, as well as draft original copy or get detailed responses to queries and even lengthier content. This is where our discussion below comes in: tools like QuillBot and others are being used to create content that is being published, purposed for the web, and much more.

Related Post: Inbox Overload? How ChatGPT Can Help Manage Your Email Account


Is AI Generated Content A Problem?

First of all, is it even an issue to use AI in writing? What are the potential problems that can come from publishing content produced by artificial intelligence? There are pros and cons to any technology or business process, and AI is no different.

Most of the people who want to avoid AI are those paying for content creation. You don’t want to shell out big bucks for a well-developed and original piece with nuanced thought and have the “writer” rely on AI. So, it makes sense to try to dissuade people you hire from writing using AI.

Other than that, you might just want to be aware of AI content for the following reasons. By looking for a few indicators (which we will go into below), you can better spot AI and avoid some of these negative consequences. If you don’t, though, here are some of the issues that can come with AI generated content:

Lack of gray area

AI tools can produce impressive results when it comes to quantifiable or accurate information. However, since it doesn’t have any degree of emotion, there is no nuance in the responses you’ll get. Algorithms can be programmed to calculate numbers and regurgitate facts, but they can’t “read between the lines”. That can make it tricky to assemble pieces that require human emotional intelligence. For example, if you ask an AI tool about a particular scenario, you’re likely to get a strictly right or wrong view, with no exploration of anything in between.

Potential plagiarism

AI content is based on information that is already in its data set - that means its out there somewhere already. For this reason, AI content can often resemble information that is currently available online. Since AI algorithms are designed to “learn” from existing data and replicate it as closely as possible, its possible you could be accused of plagiarism. The fact is that AI content is often not original or unique enough to pass copyright tests. There is a distinct possibility that your reader might see your content and think “I swear I’ve seen this before.” Not only that, if you are creating content for business, and your competitors use AI for similar topics, you’re likely to end up with content that mirrors theirs. For example, if you’re a marketing agency and wish to create a blog about The ROI on Facebook Advertising - and a competing firm wishes to write on the same topic - and you both use AI, the material will be remarkably similar.

Limited language capabilities

AI is, after all, based on machines - and not humans. AI models have a limited understanding of real language and its nuances. That means the tools likely need more help when working on complex tasks or more creative pieces. AI is great for coming up with factual or technical information, but not so great at communicating in a more complicated way.

Frowned on in search engines

Google has since clarified that they do not in fact penalize content created by AI - but why take the risk? And make no mistake there is a risk involved. That’s because AI content tends to be redundant and learn from existing content. Since search engines (and Google in particular) penalize a lack of authenticity and reward credible content by ranking sites higher, filling your website with material that could be found elsewhere won’t do much to help you. Essentially, SEO experts will tell you that the more original and high-quality content you can publish on your website, the better. If you use AI and fill your website with information that is replicated elsewhere, search engines are likely to ding you.

5 Ways to Detect AI Writing

Alright, so there are some great things about AI, and some less than great things. If you are requesting, requiring, or paying for any sort of content creation, you might want to keep an eye out for AI-generated content. It’s worth noting that the rise in AI has led to the development of AI detection tools. Several different options claim that they can tell you if content was developed by AI - but proceed with caution. Most experts agree that these tools are neither reliable nor accurate. They often flag content that was not in fact created by AI, and let some slip through that were in actuality written by AI. While these tools can be helpful in identifying patterns and raising concerns, they really can’t be relied on to pinpoint AI content.

Rather, people who are responsible for reading and judging content should look for the following indicators which might raise a red flag about how the content was generated. Experts suggest that you focus on the following:

1. Tone and style

With AI, the tone and style can be really inconsistent. They continue to make progress and improve, but for now they can struggle to produce human-like verbiage. You’ll often see output that is strung together without transition words, varying tones, or normal word placement. With human-created content, you often see varying tones throughout the text with a lot of transition words. Computer-generated content often has a more uniform tone in sentence after sentence.

2. Lack of accuracy

ChatGPT was trained on data through 2021. Therefore, these tools are not a great source of information for more current or recent events. The information you obtain can be really skewed if the most accurate data is from the past few years. Additionally, we all know the internet is not without fault. Have you heard the expression “You can’t believe everything you read online?” That’s very good advice, and for good reason. That means you need to fact-check everything that AI puts forth. Though chances are that the information you get is accurate, you never know. Be sure to fact-check the information presented in content. If it’s not verifiable in other sources, it could be AI  generated.

3. Lack of personal touches

Where is the emotion? What is the opinion put forth in the piece? If you can’t find either of those, then the content is probably AI generated. That’s perfectly fine for academic or technical content, but not so effective for think-pieces or thought leadership. If you can’t find any emotion or personal beliefs shared (in a piece where such is appropriate), then it might be AI.

4. Repetitive language

This is a big one when it comes to spotting AI! Keyword stuffing used to be a common practice in SEO, but now digital marketers know better. If you notice the same words - particularly those that could be considered keywords - used over and over, that’s a strong indicator of AI. That’s because people often use keywords in their prompt, and so the data that AI tools use will be sure to include the relevant information wherever possible. All of this generally leads to seeing the same terms over and over in text, which isn’t good for SEO or for readability. Another thing that tends to happen with AI text is the same word being used to start multiple sentence. Most people, when writing, vary their terminology and try to avoid being repetitive. AI on the other hand, will seek to include the prompted term several times, which can lead to several sentences beginning with that word. As an example, if someone writes an essay on George Washington, you’re likely to see several sentences begin with “George Washington” instead of “He” or “The president”.

5. Too many of certain words

Keep your eyes peeled for an overuse of “the”, “it”, and “is”. That’s because language models work by predicting the next word in a sentence. They are much more likely to use these words than others that are more rare or even odd. This is one of the big indicators that automated detector systems are good at picking up. It’s also good to know that these words (instead of more complex ones) can make text look really crisp and clean - which might lead you to believe the content was created by humans. If something reads a little too cleanly, there is an additional chance that AI is in play.

Related Post: How to Use ChatGPT to Create a Resume


Should I Avoid AI Altogether, Then?

The truth is, as good as AI has gotten, there is no substitute for unique, original content. Whether you’re writing a blog post, an essay response, or a recipe, your own thoughts and expressions are the way to go. That being said, it is a good idea - and in fact, a smart one - to use AI programs to your advantage. When developing your own unique content, don’t be afraid to employ AI for the following tasks:

  • Content ideas - Use AI to come up with ideas for your next blog post, social media caption, product descriptions, etc. You can also use it to brainstorm things like titles, heading options, etc.
  • Audience analysis - AI can assist content creators in understanding their audience better. Use the tools to analyze audience demographics, behavior, and engagement patterns.
  • Keyword research - Though you want to avoid keyword stuffing, keyword research is essential for SEO efforts. You can begin your research by asking AI tools what the most popular terms are for certain searches, or which phrases are most inquired on any given topic.
  • Video and image editing - AI can automate and streamline the entire video and image editing process. Some tools can automatically enhance images, remove backgrounds, apply filters, and even generate summaries for publishing.
  • Translation or localization - AI-powered translation tools can help you to reach a more global audience by translating current content into other languages. Some tools can also assist you in localizing content by finding cultural nuances and preferences.
  • Content optimization - Optimize content by analyzing and obtaining insights that can help your content performance. You can also use AI tools to review your piece and make suggestions for improvement, from grammar to SEO optimizations.

Artificial intelligence is an extremely relevant and fascinating topic, and one we’ve covered a lot on our blog - and will continue to. Make sure to follow along to get our take on all things technology, from trends to the latest tips.