Book a Call
AI-Generated Content
AI content detection
Human vs AI content
AI content patterns
AI detectors
ARTICLE #20
Table of contents
How to detect AI-generated content? Manually and with tools
AI-Generated Content
AI content detection
Human vs AI content
AI content patterns
AI detectors
AI-Generated Content
AI content detection
Human vs AI content
AI content patterns
AI detectors
Written by:
5 min read
Updated on: March 15, 2024
Toni Hukkanen
Head of Design
Creative Direction, Brand Direction
Toni Hukkanen
Head of Design
Creative Direction, Brand Direction
Toni Hukkanen
Head of Design
Creative Direction, Brand Direction
Almost every day, a new story circulates about the delights and perils of AI. When it comes to AI apps, the two major players adopted by the masses are AI content generators and AI detectors.
Many have started questioning the authenticity of the content they read and want to know if a human or AI wrote it. You might also have tried ChatGPT or other AI writing tools and noticed that the text tends to be more generic and predictable.
There are plenty of AI detectors available in the market that aren't 100% accurate because they operate using language models similar to those powering tools like Claude or ChatGPT. Also, these tools are not substitutes for critical thinking, and that's when you need to combine them with your own judgment while paying attention to the details. Keep reading to determine how to detect AI content manually or using tested AI detectors.
Almost every day, a new story circulates about the delights and perils of AI. When it comes to AI apps, the two major players adopted by the masses are AI content generators and AI detectors.
Many have started questioning the authenticity of the content they read and want to know if a human or AI wrote it. You might also have tried ChatGPT or other AI writing tools and noticed that the text tends to be more generic and predictable.
There are plenty of AI detectors available in the market that aren't 100% accurate because they operate using language models similar to those powering tools like Claude or ChatGPT. Also, these tools are not substitutes for critical thinking, and that's when you need to combine them with your own judgment while paying attention to the details. Keep reading to determine how to detect AI content manually or using tested AI detectors.
Incorrect and outdated information
Incorrect and outdated information
Most of the time, AI content may look professional and well-written, but in reality, it could be irrelevant, misleading, or just hallucinations.
If you clearly notice factual errors, outdated information, or misinterpreted context, know that the content is likely generated by AI rather than a human.
Freelance writers, copywriters, and contractors often use AI tools for an article for a company's blog. Statistics show that over 15% of writers use an AI tool to enhance their writing. But, an AI-generated article will be inaccurate and probably similar to everything already posted or plagiarised.
Businesses want to publish only relevant, well-optimised, and updated content with a unique point of view. If their writers use AI tools to generate posts, this could badly harm the brand and waste company resources.
Most of the time, AI content may look professional and well-written, but in reality, it could be irrelevant, misleading, or just hallucinations.
If you clearly notice factual errors, outdated information, or misinterpreted context, know that the content is likely generated by AI rather than a human.
Freelance writers, copywriters, and contractors often use AI tools for an article for a company's blog. Statistics show that over 15% of writers use an AI tool to enhance their writing. But, an AI-generated article will be inaccurate and probably similar to everything already posted or plagiarised.
Businesses want to publish only relevant, well-optimised, and updated content with a unique point of view. If their writers use AI tools to generate posts, this could badly harm the brand and waste company resources.
Lack of depth and personal touch
AI tools are designed to replicate existing information without adding new perspectives or insights. This intrinsic limitation leads to quick content creation, which often lacks the depth and originality that human writers bring to the table.
These tools do not really write but generate text based on patterns in their training data, which is why they don't understand what they are writing about in the same way humans do. The results are often superficial and shallow responses without critical thinking and deep analysis.
Artificial intelligence doesn't have a personality, so AI-generated content will lack a personal touch and sound emotionless and robotic.
On the other hand, a copywriter or a journalist has real conversations with subject matter experts to gain a deeper understanding, interesting stories, and relevant opinions that can't be replicated with AI tools.
AI tools are designed to replicate existing information without adding new perspectives or insights. This intrinsic limitation leads to quick content creation, which often lacks the depth and originality that human writers bring to the table.
These tools do not really write but generate text based on patterns in their training data, which is why they don't understand what they are writing about in the same way humans do. The results are often superficial and shallow responses without critical thinking and deep analysis.
Artificial intelligence doesn't have a personality, so AI-generated content will lack a personal touch and sound emotionless and robotic.
On the other hand, a copywriter or a journalist has real conversations with subject matter experts to gain a deeper understanding, interesting stories, and relevant opinions that can't be replicated with AI tools.
Overly formal or mechanical language
ChatGPT often overuses some words like "utilise," "implement," "ascertain," "leverage," and "elucidate," but will humans use these words when writing a general article? Probably not, because simple terms like ‘use, explain, and find’ in human conversations are more common and relatable.
Human writing also includes the use of slang and idioms, but when AI tries to learn from the way we talk in an attempt to copy our style, it often gets those slightly wrong. If you notice an idiom or slang in content that feels forced, know that AI has played its part.
Other favourite words of ChatGPT are "harness," "unveiling the power," "realm," "demystify," "unravel," "beacon," "unleash," "dive," "delve," "elevate," "unlock," "tailor," "ever-evolving," "landscape," "meticulously," "navigate," "embark," "tapestry, "paradigm," "resonate," and "discover."
Note: Not every piece of content with words like dive, delve, or discover should be suspected as AI-generated content; as these tools are trained using human-generated content and data, they will devour every word humans use. So, count on other factors as well to manually check if a content piece is AI-generated or not.
ChatGPT often overuses some words like "utilise," "implement," "ascertain," "leverage," and "elucidate," but will humans use these words when writing a general article? Probably not, because simple terms like ‘use, explain, and find’ in human conversations are more common and relatable.
Human writing also includes the use of slang and idioms, but when AI tries to learn from the way we talk in an attempt to copy our style, it often gets those slightly wrong. If you notice an idiom or slang in content that feels forced, know that AI has played its part.
Other favourite words of ChatGPT are "harness," "unveiling the power," "realm," "demystify," "unravel," "beacon," "unleash," "dive," "delve," "elevate," "unlock," "tailor," "ever-evolving," "landscape," "meticulously," "navigate," "embark," "tapestry, "paradigm," "resonate," and "discover."
Note: Not every piece of content with words like dive, delve, or discover should be suspected as AI-generated content; as these tools are trained using human-generated content and data, they will devour every word humans use. So, count on other factors as well to manually check if a content piece is AI-generated or not.
Predictable patterns (e.g., common phrases)
AI is neither perfect nor smart; it is simply learning from human insights and trying to replicate whatever it has learned. If you are reading content with similar phrases or the same words being used repeatedly, know that it's AI-generated.
There are other signs as well, like too many emojis on social media posts, especially the rocket emoji 🚀, starting a long-form copy with some variation of "in today's digital landscape" and alliterations to create an appealing arrangement of words.
Some spammy AI generation SEO tools prioritise keyword stuffing, with some articles having their target keyword in every other sentence. This pattern is extremely off-putting for readers.
ChatGPT also loves transitional words like "furthermore," "moreover," "consequently," "additionally," and "hence" after every few lines, which isn't very common in human writing. Human content doesn't include transitions unless it's professional or formal.
AI is neither perfect nor smart; it is simply learning from human insights and trying to replicate whatever it has learned. If you are reading content with similar phrases or the same words being used repeatedly, know that it's AI-generated.
There are other signs as well, like too many emojis on social media posts, especially the rocket emoji 🚀, starting a long-form copy with some variation of "in today's digital landscape" and alliterations to create an appealing arrangement of words.
Some spammy AI generation SEO tools prioritise keyword stuffing, with some articles having their target keyword in every other sentence. This pattern is extremely off-putting for readers.
ChatGPT also loves transitional words like "furthermore," "moreover," "consequently," "additionally," and "hence" after every few lines, which isn't very common in human writing. Human content doesn't include transitions unless it's professional or formal.
Superficial treatment of complex topics
Another way to tell whether a piece of content is AI or human-written is to check for a lack of complex analysis. Machines are undoubtedly great at data collection but are not good at turning that data into something meaningful.
If you read a blog post filled with facts without real insights and analysis, there is a high chance it was generated with AI. ChatGPT might not be smart, but it's really good at generating false or imaginary facts and figures.
You will also notice that AI-generated content is much better for static writing ( facts, history, etc.) than analytical writing. As AI can't think critically, question assumptions, or generate original insights, it just starts replicating patterns from its data without the ability to incorporate genuine problem-solving and abstract reasoning to handle complex topics.
Another way to tell whether a piece of content is AI or human-written is to check for a lack of complex analysis. Machines are undoubtedly great at data collection but are not good at turning that data into something meaningful.
If you read a blog post filled with facts without real insights and analysis, there is a high chance it was generated with AI. ChatGPT might not be smart, but it's really good at generating false or imaginary facts and figures.
You will also notice that AI-generated content is much better for static writing ( facts, history, etc.) than analytical writing. As AI can't think critically, question assumptions, or generate original insights, it just starts replicating patterns from its data without the ability to incorporate genuine problem-solving and abstract reasoning to handle complex topics.
Lack of credible references and citations
ChatGPT often inserts unreliable, incorrectly linked, and improperly formatted citations. If you spot such citations in content, it'll be clear to you that the content is AI-generated. Though AI can process vast amounts of historical data, it struggles to keep up with recent developments or new research. Without access to current information databases or direct links to credible sources, AI tools provide outdated information without proper attribution to up-to-date sources.
Using AI tools without oversight or fact-checking delivers inaccurate information that lacks credible references and citations, which can damage your credibility in your industry.
ChatGPT often inserts unreliable, incorrectly linked, and improperly formatted citations. If you spot such citations in content, it'll be clear to you that the content is AI-generated. Though AI can process vast amounts of historical data, it struggles to keep up with recent developments or new research. Without access to current information databases or direct links to credible sources, AI tools provide outdated information without proper attribution to up-to-date sources.
Using AI tools without oversight or fact-checking delivers inaccurate information that lacks credible references and citations, which can damage your credibility in your industry.
5 tested AI content detectors that are getting better every day
The most straightforward way to detect AI-generated content is to pay close attention. However, there are plenty of AI detectors online that you can use to check if a post is written by a human or AI. There are free and paid tools, but the authenticity of each varies. That's why we have tried and tested multiple tools and listed these 5 that are improving daily. However, be cautious before relying on the results of any or all of these tools.
1. Writer.com AI Content Detector
Writer.com is an AI-generated content service provider focused on corporate teams. Its AI Content Detector scans content once you add a URL or copy and paste the text into the window. After hitting the Analyse Text button, the platform provides a percentage of AI-generated copy detected in the article.
Pros
It doesn't require a sign-up to use an AI detector. Teams can even use a shared membership. It also provides instant results.
Cons
It doesn't highlight the portion of text suspected to be AI, and its results are usually undependable.
2. GPTZero
GPTZero analyses texts based on their perplexity and burstiness. Perplexity is the randomness of a text to a model or how well a language likes a text. Burstiness is how uniform the perplexity is over time compared to varied human-written text.
It's a place where you can register for a free account to scan over 5,000 words without login. If you want to detect GPT content, the response you'll get will have more details and increased accuracy.
Pros
Like other AI detectors, it is free to use. You can detect GPT-2 text with this tool without signing up.
Cons
It lacks a lot of features like a plagiarism checker, fact checker, and readability checker.
3. ZeroGPT
ZeroGPT detects whether a given text is generated by an AI tool like Google Bard, ChatGPT, or a human. This tool uses DeepAnalyse technology to determine the origin of text. ZeroGPT claims an accuracy rate of over 98% and can detect AI text in all available languages.
Pros
The accuracy rate is higher in ZeroGPT, which has analysed over 10M articles and texts.
This tool is straightforward to use. All you need to do is enter text in the box and click Detect Text to get the results (the percentage and highlighted AI-generated text).
The best thing about ZeroGPT is privacy, as your text won't be saved or available online after this AI detector detects it.
Cons
There are no features to proofread and edit content after detection to increase the content quality.
Even though it has high accuracy, in the end, it's a tool that's not perfect and can generate false positives and false negatives.
4. Originality.ai
Originality AI is a great tool for anyone who wants to detect AI-generated text. Though its plagiarism checker and readability scanner work wonders, they sometimes misidentify human-written content. As no AI detector is perfect, Originality AI usually gives the right results.
Pros
The results are usually accurate. Due to flexible pricing, you can even use its paid version. A free tool on the website is available to test it.
Cons
It only supports English, lacks in-depth results, and fails to detect more advanced humanisers like undetectable.ai.
5. QuillBot
QuillBot is an AI paraphrasing tool that also offers its own built-in AI detection software. This software claims to detect AI content from a variety of AI writing tools such as GPT-4, ChatGPT, and Google Gemini.
Pros
You can check AI without signing up. There are extensions available for Microsoft Word, macOS, and Google Chrome. You can also access a free language translator.
Cons
The results vary from time to time, with accuracy between 45% to 80%.
The most straightforward way to detect AI-generated content is to pay close attention. However, there are plenty of AI detectors online that you can use to check if a post is written by a human or AI. There are free and paid tools, but the authenticity of each varies. That's why we have tried and tested multiple tools and listed these 5 that are improving daily. However, be cautious before relying on the results of any or all of these tools.
1. Writer.com AI Content Detector
Writer.com is an AI-generated content service provider focused on corporate teams. Its AI Content Detector scans content once you add a URL or copy and paste the text into the window. After hitting the Analyse Text button, the platform provides a percentage of AI-generated copy detected in the article.
Pros
It doesn't require a sign-up to use an AI detector. Teams can even use a shared membership. It also provides instant results.
Cons
It doesn't highlight the portion of text suspected to be AI, and its results are usually undependable.
2. GPTZero
GPTZero analyses texts based on their perplexity and burstiness. Perplexity is the randomness of a text to a model or how well a language likes a text. Burstiness is how uniform the perplexity is over time compared to varied human-written text.
It's a place where you can register for a free account to scan over 5,000 words without login. If you want to detect GPT content, the response you'll get will have more details and increased accuracy.
Pros
Like other AI detectors, it is free to use. You can detect GPT-2 text with this tool without signing up.
Cons
It lacks a lot of features like a plagiarism checker, fact checker, and readability checker.
3. ZeroGPT
ZeroGPT detects whether a given text is generated by an AI tool like Google Bard, ChatGPT, or a human. This tool uses DeepAnalyse technology to determine the origin of text. ZeroGPT claims an accuracy rate of over 98% and can detect AI text in all available languages.
Pros
The accuracy rate is higher in ZeroGPT, which has analysed over 10M articles and texts.
This tool is straightforward to use. All you need to do is enter text in the box and click Detect Text to get the results (the percentage and highlighted AI-generated text).
The best thing about ZeroGPT is privacy, as your text won't be saved or available online after this AI detector detects it.
Cons
There are no features to proofread and edit content after detection to increase the content quality.
Even though it has high accuracy, in the end, it's a tool that's not perfect and can generate false positives and false negatives.
4. Originality.ai
Originality AI is a great tool for anyone who wants to detect AI-generated text. Though its plagiarism checker and readability scanner work wonders, they sometimes misidentify human-written content. As no AI detector is perfect, Originality AI usually gives the right results.
Pros
The results are usually accurate. Due to flexible pricing, you can even use its paid version. A free tool on the website is available to test it.
Cons
It only supports English, lacks in-depth results, and fails to detect more advanced humanisers like undetectable.ai.
5. QuillBot
QuillBot is an AI paraphrasing tool that also offers its own built-in AI detection software. This software claims to detect AI content from a variety of AI writing tools such as GPT-4, ChatGPT, and Google Gemini.
Pros
You can check AI without signing up. There are extensions available for Microsoft Word, macOS, and Google Chrome. You can also access a free language translator.
Cons
The results vary from time to time, with accuracy between 45% to 80%.
Is AI-generated content bad for SEO?
Businesses often fear that AI content doesn't rank well on search engines and is bad for SEO. It simply means their marketing budget is wasted if their writers use AI to generate site content.
Currently, Google has made it clear that AI-generated content will not impact search rankings as long as your content is original, helpful, and relevant. Google is also less concerned about how you generate content and more concerned with the quality of the content.
Using the EEAT framework, Google continues to reward helpful content that demonstrates expertise, is published on authoritative websites, and is trustworthy.
Google also mentions in its long-standing spam policy that the use of automation, including generative AI, is spam if the main purpose is to manipulate ranking in search results.
Businesses often fear that AI content doesn't rank well on search engines and is bad for SEO. It simply means their marketing budget is wasted if their writers use AI to generate site content.
Currently, Google has made it clear that AI-generated content will not impact search rankings as long as your content is original, helpful, and relevant. Google is also less concerned about how you generate content and more concerned with the quality of the content.
Using the EEAT framework, Google continues to reward helpful content that demonstrates expertise, is published on authoritative websites, and is trustworthy.
Google also mentions in its long-standing spam policy that the use of automation, including generative AI, is spam if the main purpose is to manipulate ranking in search results.
Frequently Asked Questions
Can Google identify AI-generated content?
Yes, even though the content is accurate, Google can detect AI. This tech giant employs various methods to sift through content that might contain AI.
How to pass the AI content detector?
You can avoid AI detection if you rephrase sentences using different words while keeping the original meaning there. You should avoid monotony and use varying sentence length and structure to keep content engaging. If there are simple ideas, use short sentences to convey key points and longer and complex sentences to add detail.
Does Google penalise AI content in 2024?
Google penalises low-quality, spammy, or misleading content even if it's generated by an AI tool or a human. Sometimes, Google ranks AI content if it meets EEAT guidelines and is original, informative, and well-written.
Final Thoughts
Though it is not easy to tell if a content piece is written by AI or a human, you can combine tools and your judgment for a better assessment. However, remember to take the results you see with a grain of salt. Nothing you notice is conclusive in any way since there is no concrete way to detect AI. What you are most probably working with leaves no watermark, and you are just looking at words on a screen. AI detectors will help us filter out AI-generated content on the internet, school systems and in the news worldwide.
While AI content is undoubtedly getting better and more human-like, it still needs our attention to ensure the final draft is original and has that personal touch that makes a story stand out.
Frequently Asked Questions
Can Google identify AI-generated content?
Yes, even though the content is accurate, Google can detect AI. This tech giant employs various methods to sift through content that might contain AI.
How to pass the AI content detector?
You can avoid AI detection if you rephrase sentences using different words while keeping the original meaning there. You should avoid monotony and use varying sentence length and structure to keep content engaging. If there are simple ideas, use short sentences to convey key points and longer and complex sentences to add detail.
Does Google penalise AI content in 2024?
Google penalises low-quality, spammy, or misleading content even if it's generated by an AI tool or a human. Sometimes, Google ranks AI content if it meets EEAT guidelines and is original, informative, and well-written.
Final Thoughts
Though it is not easy to tell if a content piece is written by AI or a human, you can combine tools and your judgment for a better assessment. However, remember to take the results you see with a grain of salt. Nothing you notice is conclusive in any way since there is no concrete way to detect AI. What you are most probably working with leaves no watermark, and you are just looking at words on a screen. AI detectors will help us filter out AI-generated content on the internet, school systems and in the news worldwide.
While AI content is undoubtedly getting better and more human-like, it still needs our attention to ensure the final draft is original and has that personal touch that makes a story stand out.
Frequently Asked Questions
Can Google identify AI-generated content?
Yes, even though the content is accurate, Google can detect AI. This tech giant employs various methods to sift through content that might contain AI.
How to pass the AI content detector?
You can avoid AI detection if you rephrase sentences using different words while keeping the original meaning there. You should avoid monotony and use varying sentence length and structure to keep content engaging. If there are simple ideas, use short sentences to convey key points and longer and complex sentences to add detail.
Does Google penalise AI content in 2024?
Google penalises low-quality, spammy, or misleading content even if it's generated by an AI tool or a human. Sometimes, Google ranks AI content if it meets EEAT guidelines and is original, informative, and well-written.
Final Thoughts
Though it is not easy to tell if a content piece is written by AI or a human, you can combine tools and your judgment for a better assessment. However, remember to take the results you see with a grain of salt. Nothing you notice is conclusive in any way since there is no concrete way to detect AI. What you are most probably working with leaves no watermark, and you are just looking at words on a screen. AI detectors will help us filter out AI-generated content on the internet, school systems and in the news worldwide.
While AI content is undoubtedly getting better and more human-like, it still needs our attention to ensure the final draft is original and has that personal touch that makes a story stand out.
More news
Work with us
Click to copy
work@for.co
FOR® Industries
- FOR® Brand. FOR® Future.
We’re remote-first — with strategic global hubs
Click to copy
Helsinki, FIN
info@for.fi
Click to copy
New York, NY
ny@for.co
Click to copy
Miami, FL
mia@for.co
Click to copy
Dubai, UAE
uae@for.co
Click to copy
Kyiv, UA
kyiv@for.co
Click to copy
Lagos, NG
lagos@for.ng
Copyright © 2024 FOR®
Work with us
Click to copy
work@for.co
FOR® Industries
We’re remote-first — with strategic global hubs
Click to copy
Helsinki, FIN
hel@for.co
Click to copy
New York, NY
ny@for.co
Click to copy
Miami, FL
mia@for.co
Click to copy
Dubai, UAE
uae@for.co
Click to copy
Kyiv, UA
kyiv@for.co
Click to copy
Lagos, NG
lagos@for.ng
Copyright © 2024 FOR®