How to Spot a Blog Post Written by Chat GPT

These are strange times we are living in…

The pace at which technology is developing is terrifyingly exciting. While some movies predicted that we would all transport ourselves around by hoverboard or flying car, we seem happy to trade those lofty heights for seemingly sentient chatbots, mobile phones that hold all of our personal and financial information, and surveillance systems that track our every movement.

A bleaker depiction of artificial intelligence is Charlie Brooker’s television series ‘Black Mirror’.

The show was critically acclaimed across its five-season run, showing the very worst possible scenarios of rapidly developing technology alongside the weaknesses and evils of human beings. Episodes covered topics such as a person’s electronically stored memories ruining their marriage, and people rating each other online during social interactions to determine their standing in society.

While some may be frightened of the possibilities of artificial intelligence and some are intrigued, it seems the most dangerous thing that is happening at the moment (that we know of) is that students may be using it to write their assignments, and marketing officers may be using it to write blogs for their College’s website.

Now, while the quality of copy from Chat GPT’s responses is very well put together, it might be morally wrong, or at least morally ambiguous to use it in an educational or professional setting. It will definitely get you docked marks if you use it for an essay or assignment!

Even if every sentence is grammatically correct and the chatbot gets its point across in a sound manner, people might not appreciate a company entrusting their audience to artificial intelligence.

Here are some things to look out for:

1. Background Information – AI Chatbots love to begin a blog with background information about the topic at hand. They do this to introduce the reader to the broader topic before honing in on the main point.
2. Structure – AI Chatbots love to stick to the same structure no matter the topic. They will begin with the background before focusing on one main topic or a few main topics, keeping the reader engaged at all times.
3. Titles – AI Chatbots like to suggest a title for a piece. They like to make them attention-grabbing to hook readers from the very start.
4. Questions – AI Chatbots love to ask questions, especially at the beginning of their blogs. These will usually be rhetorical and are just to serve the purpose of the blog.
5. Lists – AI Chatbots love to utilise lists. They usually keep their blogs short and snappy, and in my experience keep it consistent at five bullet points.

Of course, by the time I have written this piece, Chat GPT and other AI chatbots may have developed even further and this blog will be obsolete. Perhaps this particular blog was written by AI and edited to make you believe I wrote it. Maybe Cian Dalton is the latest entirely artificially ‘intelligent’ marketing officer to hit the market, pardon the pun.

So, keep an eye out for these signs and let me know if you spot any fake blogs out there!

For reference, here is the response from Chat GPT when asked, “Can you tell me how to know if a blog was written by an artificially intelligent chatbot?”

“It can be challenging to determine with certainty whether a blog post was written by an artificially intelligent chatbot, as some AI-generated content can be highly sophisticated and difficult to distinguish from human writing. However, there are a few clues you can look for that may indicate the content was produced by AI:

1. Lack of personality: AI-generated content is often characterized by a lack of personality or voice. It may be well-written and informative, but it will lack the personal touch and style that human writers bring to their work.
2. Repetition: Chatbots are programmed to follow certain patterns and generate content based on certain rules. As a result, their writing may contain repetitive or formulaic language that is not typical of human writing.
3. Errors or inconsistencies: While AI-generated content can be highly sophisticated, it is not always perfect. Look for errors or inconsistencies in the writing that may suggest it was produced by a machine rather than a human.
4. Strange phrasing or word choices: AI-generated content may use phrasing or word choices that seem unusual or awkward to a human reader.

Overall, it can be difficult to definitively determine whether a blog post was written by an AI chatbot or a human writer. However, by paying attention to the factors listed above, you may be able to make an educated guess as to the source of the content.”

I found it interesting that after writing my piece, Chat GPT reduced the list to four bullet points, did not suggest any title and kept a more serious tone than other times I had used the software. 

I also asked if it “had the capacity for annoyance when asked about AI detection.” The response is below:

“As an artificial intelligence language model, I don’t have the capacity to feel emotions like annoyance or frustration. I’m here to assist and answer your questions to the best of my abilities, regardless of the topic. So, feel free to ask me anything related to AI detection, and I’ll do my best to provide you with helpful and informative answers.”

Strange times, indeed!