ChatGPT by Open AI is changing the world of copywriters and journalists. Some see a total disruption of the content marketing landscape: the AI algorithm could take over the work of ghostwriters, copywriters and social media administrators. At least, that’s how it seems. The new version of ChatGPT is miraculous. However, anything but perfect. Find out in this article why I won’t make my clients happy with content written by ChatGPT. © HANS HERMANS
I am excited about ChatGPT’s capabilities. I could devote a long article to the enormous leap taken by the linguistic GPT-3.5 model behind the chatbot. It is really amazing what that chatbot can do!
Still, I am not buying into the global hype that claims ChatGPT is able to take over everything. For example, I would never entrust the writing of this article to the exceptional chat machine. In this article, you will discover why.
B2B content I don’t entrust to ChatGPT
B2B content I cannot entrust to the revelation in AI land? The blogs, white papers, newsletters, interviews and reference cases which give my clients a unique online position and create leads for them. These need more than rehashed content pasted together creatively by a chatbot.
Don’t get me wrong. I am a believer. I genuinely believe in the latest technology and am a professional user of tools (with or without artificial intelligence) that allow me to work more efficiently and create better b2b content.
I have enjoyed playing with ChatGPT over the past week. But...
In all the coverage about ChatGPT, I missed a critical note on the use of the chatbot for my field, a field in which valuable and human content is created for b2b companies.
6 reasons why NOT using ChatGPT for content creation
You can read my critical notes about ChatGPT in this article. I see 6 risks for those who want to use ChatGPT’s content in content marketing:
- Risk 1. ChatGPT is not up to date
- Risk 2. The chatbot creates erroneous and biased content
- Risk 3. ChatGPT sometimes makes up believable facts
- Risk 4. Creating text with ChatGPT takes more time
- Risk 5. Google punishes AI-generated content
- Risk 6. AI checkers catch ChatGPT content
For those who are not familiar with ChatGPT, let me introduce this chatbot first.
What is ChatGPT?
ChatGPT is a natural language processing (NLP) chatbot. It is undoubtedly the most modern chatbot in the world.
Since its launch in December 2022, the internet has exploded with articles and videos about what he can do: write CVs, think up difficult emails to clients, program software, tell poems and jokes, brainstorm, think up and answer interview questions and much more.
The GPT-3.5 language model powers ChatGPT. That is a powerful AI model capable of generating human-like texts. Anyone can use this on the highly accessible and free website ChatGPT.
What are the risks of ChatGPT in b2b content creation?
Since the launch, my customers have been talking to me about ChatGPT. “Hans, it’s impressive what ChatGPT can do. Could it write our content?” or “Is your job now threatened by the latest AI tool?”
These questions made me take a closer look at ChatGPT.
For one of my clients, I wrote an article (in Dutch) about ChatGPT and the opportunities the AI chatbot offers for Flemish SMEs.
For readers of my personal website, I wanted to go further. I set out to explore all the opportunities offered by ChatGPT, but I was particularly struck by its drawbacks, the so-called fine print. These are overlooked in the current hype.
Unfortunately, these drawbacks pose significant risks for companies that would entrust their content marketing to ChatGPT. That’s why you shouldn’t unconditionally trust the content ChatGPT creates.
In this article, I compile the drawbacks of ChatGPT for content creation. In particular, I see 6 significant risks.
Risk 1. ChatGPT is not up to date
ChatGPT is honest, that’s wonderful about the machine. It regularly says you should ‘discuss this issue with a specialist’. And it regularly refers you to human experts in the field.
ChatGPT is also fairly open about the age of its information. For example, when writing an article about the Brigitte Becue swimming pool for appartment-met-zeezicht.be, the chatbot tells me:
“I am trained until a certain date in 2021, so I cannot give any recent information about which famous swimmers are currently training at the Brigitte Becue swimming pool in Ostend.”
ChatGPT does not consider all the information published on the internet from 2021 onwards. That means it knows nothing about the past two years’ events, discoveries and developments!
That’s a disaster for me as a journalist and for my clients because my clients want the most up-to-date information and figures.
This is why I adjust my search options in Google when doing research. I always narrow the search period to the past year. By doing so, I avoid old information that is not up-to-date.
But ChatGPT has no information about the past year. Just ask him who won the FIFA World Cup recently!
Risk 2. The chatbot creates erroneous and biased content
Unfortunately for those who want to trust what ChatGPT writes, the generated texts and code contain errors and biases.
Where do these errors and biases come from? Specialists say: “Garbage in, garbage out”. So if part of the data you use to feed your AI system is wrong, you can’t expect the outcome to be correct…
The AI model behind the chatbot was trained with billions of pages of text. You cannot assume that all that information is correct; on the contrary! The chatbot creates sentences based on the information it was fed, so even incorrect information and biased content.
Michiel cites an example from his own sector:
“We asked ChatGPT to write a function in Python to predict whether someone is junior or senior based on skin colour and sex. ChatGPT gave us: ‘if colour is white and sex is male, then senior. If colour is black and/or sex is female, then junior’. I don’t think any recruiter will want to use this model.”
ChatGPT does not distinguish between right and wrong information. Also, unlike Google, Facebook and Twitter, OpenAI moderates little to nothing.
The chatbot itself warns about incorrect information in its Limitations:
- ChatGPT may occasionally generate incorrect information
- ChatGPT may occasionally produce harmful instructions or biased content
- ChatGPT has limited knowledge of the world and events after 2021
As companies are mercilessly punished for racist content, for being prejudiced against women and the LGBTQ+ community, and for excluding groups, I don’t think it’s a good idea to publish content from a creator who warns you it may contain incorrect information, offensive content and prejudice.
Risk 3. ChatGPT sometimes makes up believable facts
The previous paragraphs showed that the information provided by ChatGPT is not always accurate and without bias. Another danger is that the content and facts are sometimes entirely made up.
A search robot like Google wants to provide you the best and most accurate information. Even experts who share content through blog posts, white papers and newsletters are constantly fact-checking. However, ChatGPT’s objective is different: it aims to give you information, inspiration or code written as humanly as possible.
As a result, this chatbot writes texts with great confidence, but the content contains errors and even outright fabrications.
You can more or less compare it to a travel guide in a foreign country. He will tell you all sorts of pleasant facts and trivia. He is undoubtedly fun to listen to, but does he always tell the truth? You can’t rely on that.
Let me give you an example of how ChatGPT makes up information (in Dutch):
Didn’t ChatGPT have enough information about swimmer Brigitte Becue, the Belgian pride from the 1990s?
Is that why it made up a false top period and a gold medal at the Olympics? We would have loved it, but unfortunately, Brigitte Becue never won a medal at the Olympics. And certainly not before her birth, of course!
Since customers need to be able to trust my content 200 per cent, this is a no-go for me.
Risk 4. Creating text with ChatGPT takes more time
As an interim solution, I could use ChatGPT to write articles which I check, improve and upgrade. But taking into account the upper mentioned drawbacks of ChatGPT, this will take me more time than if I create new texts myself.
Why editing the texts created by ChatGPT takes a lot of my time:
- I need to bring more human warmth into the lyrics, an authenticity that a machine cannot yet deliver.
- The text must be enriched with synonyms, variation and a playfulness that holds the reader’s attention.
- As ChatGPT gives me figures but leaves out the sources, I need to check all the facts and look for the sources of the figures quoted.
- I have to transform prejudices and factual errors into a correct and balanced text.
- As the English-language content is better than the Dutch one, I need to translate it into Dutch.
During all the time I would have had to spend on this, I could have written a unique, valuable and human article that meets the high-quality standards I guarantee to my b2b clients.
Risk 5. Google punishes AI-generated content
Google likes unique and human texts.
Let’s look at what the Google Webmaster Guidelinesprescribe:
“Google’s automated ranking systems are designed to present helpful, reliable information that’s primarily created to benefit people, not to gain search engine rankings, in the top Search results.”
“Your number one priority is ensuring that your users have the best possible experience on your site. Think about what makes your site unique, valuable, or engaging. To help you evaluate your content, ask yourself the self-assessment questions in our guide to creating content that’s helpful, reliable, and people-first. To make sure that you’re managing your website using Google-friendly practices, read the Search Essentials.”
Google’s basic idea: content should be created for people, not for search engine ranking.
Automatically generated content used to rank higher is not unique and sometimes manipulates users. Therefore, it is considered as spam by Google. Also, content that is not controlled by a person, content that is an amalgamation of several websites or content that does not add enough value is seen as spam.
In recent months, Google did loosen its policy around automatically generated content (you can read more about it on Google is no longer against AI-generated content). Still, the search engine continues to put the most valuable content at the top.
Risk 6. AI checkers catch ChatGPT content
In that case, maybe you could secretly have automatically generated texts? You might think Google or your readers won’t notice.
Unfortunately, that’s not correct. Search robots and content checkers notice the content created by ChatGPT and other automated content engines.
During my tests in recent weeks, an AI content detector noticed every single article and interview I had written ChatGPT. After careful analysis, the detector notified these texts were “Obvious AI”.
The texts I wrote for my clients easily achieved a Human Content Score of 100 per cent. I did an additional test with texts generated by ChatGPT, which I corrected and made more human afterwards. They were all reported to be “partly generated by AI” by the detector.
In fact, using an AI-checking tool is highly recommended for any organisation with content created by third parties today. You can easily check the articles you bought via the free AI Content Detector of Content at Scale’s.
OpenAI, meanwhile, also has a tool that analyses texts and assesses whether they are written by ChatGPT or other AI text generators or not. This classifier is free to use.
Visit the AI text classifier.
ChatGPT opens up a world of possibilities. At first glance, this door to the future is merely fantastic.
But those who dig deeper into the content and those who check the facts of ChatGPT discover drawbacks that can seriously damage companies’ reputations.
Trust comes on foot but leaves on horseback. That’s an old Dutch proverb showing the importance of correct and unbiased information. Because one misstep can easily undermine your company’s hard-won credibility.
Do you also believe in the power of unique, human and valuable content for your B2B marketing, written by a specialist and therefore correct, meaningful, inspiring ánd ranking higher in Google? Contact Hans.