Ten telltale signs your content was written by AI

The use of AI-generated content is rapidly increasing as organisations seek innovative solutions to keep pace with the demand for engaging material. Many companies are leveraging AI tools to streamline content creation, allowing for quicker turnaround times and enhanced productivity.

If you couldn’t already tell, the paragraph above was written entirely by artificial intelligence (AI). Not only are the American spellings a dead giveaway, but terms like ‘leveraging’ and ‘seek,’ and the robotic tone doesn’t reflect how humans write or speak. Yet.

The internet has been saturated with content — good, bad and ugly – for a long time. But the launch of generative AI tools, notably ChatGPT in 2022, means we can produce content on a scale never previously seen. New generative tools like Perplexity and Jasper AI have since been launched, while ChatGPT now has access to more up-to-date information and produces human-like content.  
According to experts, by 2026 90% of online content will be AI generated making it harder to spot the difference between machine- and human-written content.

Does it matter if your content was written by AI?

This isn’t to say that AI content is bad. 

In fact, AI has proven to speed up the writing process by 40% and it’s a great starting point, especially for less confident writers to get ideas and inspiration.

But relying on it too heavily, without proper editorial oversight, could mean you end up churning out monotonous, strange or inaccurate content. 

One of the big issues with machine-generated content is its generic and monotonous writing style that’s difficult to adapt for a specific audience.

You can ask it to write in a specific voice, but the results can be unnatural. Blogs, social media and thought leadership content, in particular, work best when they’re authentic and, above all, human. 

We asked it to write a social media post in the voice of an artist: 

I’m thrilled to unveil my latest painting, “Whispers of the Dawn”! This piece captures the gentle embrace of morning light as it dances over a tranquil landscape, inviting viewers to experience the beauty of new beginnings. 

Some of the bigger brands, like Ryanair and Lidl, have distinguished themselves through social media with their conversational, humorous and creative brand voice. While AI can be used as a starting point to create or finesse a post, like the example above, it doesn’t always get it right. 

As writers, we tell stories fuelled by emotion and real-life experiences. Since AI doesn’t have that, it can only reproduce what’s on the web. We can also interview experts, draw on relevant research, and link to current events and trends.

Is AI content bad for SEO?

Content needs to be helpful and relevant to the audience to rank highly. If AI can generate that, it’s not bad for SEO in the slightest. 

Google’s helpful content recommendations outline that if AI has been used to create content, is it obvious? Did you provide a disclaimer? And have you explained why it was used? 

This is to help readers understand how the content was created. So, if you used AI, you might need to add a disclaimer to avoid losing trust from your audience. Some news agencies believe AI content should be labelled just like sponsored content, and others don’t see a problem in using AI disclaimers for less serious topics to get news published quickly. 

Ultimately, the search engine rewards high-quality content ‘however it is produced.’ So, as long as the content meets E-E-A-T (expertise, experience, authoritativeness and trustworthiness) guidelines and is helpful to your audience, it should be fine. 

The question is then whether AI can produce high-quality content to rank in the top spots. As it stands, without being adapted by a human writer, AI content could be too generic, especially when trying to capture traffic for niche keywords.

AI might produce a near perfect SEO article but equally it might not – you need to be able to tell the difference and think about what a human might enjoy reading. 

“You risk getting penalised if you are using AI to manipulate Google’s ranking systems, but using AI itself is not against any rules or guidelines. If you want to produce high-quality, unique and genuinely helpful content for your audience (stuff that actually works in SEO), this is generally best written by an expert who has creativity and experience.” 

Jake Cassedy, SEO lead at Tank

Ten signs your content has been written by AI

To check if content was written by AI, you can look for a disclaimer within the text or use an AI detection tool. Just be aware that these tools aren’t 100% accurate. You’re better off using the clues below.

Unnatural sentence structure

As writers, it’s our job to make writing flow. For that, you need a mix of long and short sentences to keep your readers engaged. 

Content generated by AI doesn’t tend to have sentence variety, which is why the copy often lacks rhythm. Instead, it typically generates a brief introduction, followed by a bullet point list and short conclusion, which is formulaic and not very exciting to read.

Repetition and lists

We’ve noticed that AI has a knack for in-sentence lists. While it’s good to provide detail, overusing lists is distracting and can make the writing feel cluttered. 

Take the example below. Is it easy to read? Does it flow well? 

Giveaway words and phrases

If AI created its own dictionary, we can say with conviction that these following phrases would be included:

• Delve
• Fosters
• Elevate
• Embrace
• Robust
• Daunting
• Immerse

It also has its favourite phrases, which are often repeated or slightly tweaked:

• In today’s fast paced world
• In the ever-evolving landscape
• In summary
• Dive into
• It’s important to note
• As a result
• Competitive edge

Bizarre claims

It’s not a good idea to blindly trust what AI generates. 

Names, dates, facts and stories should be double checked before being used, because there is a risk the information is wrong. For example, 33 physicians asked ChatGPT medical information but the results were largely inaccurate.

There have also been some odd responses, such as when Google’s AI advised users to: ‘add glue to pizza’ and “eat one rock per day.” ChatGPT has also had its moments like when it began generating responses in Spanish and repeating ‘happy listening.’

We put it to the test and asked how young people say hello. 

While the response isn’t as bizarre as glue on pizza, it’s not really reflective of society today. And if you’re trying to reach a Gen Z audience, relying on AI might mean that your content falls flat.

Lack of specificity and conviction

AI prefers to be neutral when it comes to decision making or views, using terms like ‘Some people may believe’ or ‘It’s ultimately your choice.’ 

If you’re looking for a strong opinion for thought leadership-style content, you need to speak to a human.

It’s too wordy

AI content can contain unnecessary words and detail making sentences harder to digest. 

We asked ‘Why good content is helpful’ and it generated: 

“High-quality content fosters a connection with the audience, making it a crucial component of any successful marketing strategy.” 

To put it into simpler words, “Good content helps you connect with your audience.” 

Content generated by AI should always be reviewed by an expert and simplified to make it easy to understand.

It isn’t personalised

The writing style of most AI tools is rigid and formal. It commonly uses terms like ‘One might conclude’ or ‘Choose what resonates.’ On the other hand, writers can choose a pronoun based on the audience. First person like ‘we’ or ‘I’ helps to soften and personalise content making it more engaging. 

We can also include examples to offer readers something to relate to or make information easier to understand. AI written tools include examples but these are often non-specific.

Lack of humour and identity

One thing that AI hasn’t got down is humour. We asked it to tell us a joke:

Because AI has no feelings or personal experiences, it can’t be funny or take on a personality, which is essential to developing a brand voice. Humans can have natural wit or sarcasm and, when carefully used within content, it can help businesses stand out in the market.

It doesn’t match your brand’s tone of voice

Big brands spend a lot of time developing their brand voice and identity. For example, language-app Duolingo is well known for its mascot’s brand voice, which it describes as expressive, playful, embracing and worldly – qualities that ultimately engage readers.

AI content generally isn’t warm, creative, humorous or empathetic, which makes it bland and uninspiring.

“With a clear set of guidelines and prompts, I could see AI being used to create consistency in a brand’s voice, perhaps for more ‘everyday’ social media or blog content. But that’s probably the limit – at least for the moment. AI copy that tries to emulate a brand voice is likely to lack warmth, empathy and creativity – and audiences will spot it a mile off.”  

Catherine Allen, head of content at Tank

American spellings

Another caveat is that most AI tools use American spellings. Watch out for the following words:

• Organizations
• Maximize
• Analyze
• Centers
• Likable
• Color
• Humor
• Labor

It’s best to ask it to respond in British English and make sure it’s been checked before publishing. 

While AI is a handy tool, it doesn’t have all of the qualities a human writer does. Overusing it can take away your brand voice and personality, and could lead to errors or misinformation.

At Tank, our writers know how to turn your expertise into campaigns that deliver results. Get in touch to find out more.