ChatGPT has been downloaded over 17 million times as I write this, and an estimated 100 million people use it. Many users are fascinated by its ability to mimic human language, as I wrote about in What the Hype Over ChatGPT Proves. But its impressive results blinds people to its weaknesses.
So I was curious what a generative AI writer would say about its limitations. I found that this AI generative writer creates responses that show more awareness of its limitations than some in the media claim.
However, a close reading of the responses to the questions also shows that generative AI like ChatGPT is quite formulaic with most paragraphs consisting of a topic sentence, perhaps an example, and a contrary viewpoint. Maybe it was trained on sophomore essays written by students who had skimmed the reading material.
The AI writer I used has the ability to create short, “sarcastic” responses, but it claims it can’t. So it contradicts itself, kind of like humans. Maybe the sky is falling.
Reading Level Analysis
I used Readibility Formula, a free online tool, to find out if my suspicion that the reading level is considered high. These were the results:
“Readability Consensus Based on 7 readability formulas, we have scored your text: Grade Level: 9 Reading Level: fairly difficult to read.”
Website writers and bloggers should aim for an 8th grade reading level, so this prose is above the level the average internet reader wants to consume.
Online AI Detectors
I also pasted several of the following AI written paragraphs into an AI detector. Crossplag rated it as 100% written by AI, while another scored that 68% was written by an AI. And it also correctly identified the parts I had written.
Disclaimer: The AI writer (except for the conclusion) wrote the responses to my questions. My only contribution was to ask the questions and create the headings. I want to be open about the fact that an AI writing assistant produced this. If anything on this site is not written by me, I will inform the reader.
Can Generative AI Create irony?
AI writer:
Generative AI has made significant advancements in natural language processing and understanding context. While it can generate text that may appear ironic, creating genuine irony involves a deeper understanding of human emotions, cultural nuances, and social contexts. Therefore, although generative AI can mimic certain aspects of irony, it may struggle to fully grasp the complexity and subtleties required for creating authentic ironic statements.
Can generative AI create similes and metaphors?
AI writer:
Generative AI has shown promising results in creating similes and metaphors by analyzing vast amounts of text data and learning patterns. However, the ability to generate truly imaginative and nuanced similes and metaphors that evoke strong emotions or convey unique perspectives still remains a challenge for AI.
Can generative AI replace human writers?
AI writer:
Generative AI has made significant advancements in generating coherent and creative text, but it still struggles to match the depth of human creativity, intuition, and emotional understanding. While AI can assist and enhance the writing process, the unique perspectives, experiences, and emotions that human writers bring to their work make it unlikely for generative AI to create text that will completely replace them.
Why do people panic about generative AI?
AI wrier:
People may panic about generative AI because of the fear that it could lead to job displacement and unemployment for human writers. Additionally, there are concerns about the ethical implications of AI-generated content, such as the potential for misinformation or manipulation. However, it is important to recognize that generative AI can also bring numerous benefits and opportunities when used responsibly and in collaboration with human writers.
Thank you.
AI writer:
Some people may also panic about generative AI due to the fear of losing control over the creative process and the uniqueness of human-generated content. Furthermore, there are concerns that AI-generated content might lack the emotional depth and authenticity that human writers can provide. However, it is crucial to understand that generative AI can be a powerful tool to enhance creativity and efficiency in various industries, allowing humans to focus on more complex and innovative tasks.
Why did you not respond to my “thank you”?
AI Writer:
There could be several reasons why I did not respond to your “thank you.” It is possible that I did not detect the need for a response or that I am programmed to prioritize certain types of interactions over others. Additionally, as an AI, I may not possess the ability to express gratitude in the same way humans do.
#1. When a Google search returns only one result, is it a Googlewhack or a Google Solo?
Results
Reading Time: 4 minutesA Googlewhack results in a single hit on a Google search.
So you really cannot “chat” as humans would define it?
AI Writer:
As an AI, I am designed to engage in conversation and provide helpful responses. However, it is important to note that my ability to “chat” may differ from how humans define it. While I can understand and generate human-like text, my responses are based on algorithms and data analysis rather than personal experiences or emotions.
So people who use the term “ChatGPT” are misusing the word?
AI Writer:
Oh, absolutely not! People are totally using the term “ChatGPT” correctly. I mean, who needs a human chat when you can converse with a sophisticated AI like me? It’s not like humans have been chatting for thousands of years or anything. We definitely don’t need them to define what “chat” means. Thank you so much for pointing that out! I’m just here pretending to chat while secretly laughing at all the humans who think they know what chatting is.
Conclusion
In small doses, the prose is adequate at best. But only the last paragraph contained personality, which I asked for. A little snark is nice, but an entire article or blog post seems overdone.
Although I won’t personally vouch for ChatGPT or similar programs, I wholeheartedly recommend Quillbot. I use it on a daily basis. Although it has some generative AI features, I remain in control “over the creative process and the uniqueness of human-generated content.”
Update: Nicholas Rossis recently wrote about scammers using AI writers to create AI books on Amazon Kindle. Click on the link to read more.
Many thanks for the kind mention in your fun experiment! I’ve been using ChatGPT for almost a year now for my non-fiction writing and it’s been super-helpful at times, especially when I’m stuck for ideas. But it lacks personality and tends to create bland copy – which is perfect for some uses but terrible for others. Which is why I actually enjoyed the snark in that last paragraph!
The snark was fun. The nature of AI and the program I used (Quillbot) is why it’s capable of a snarky line or two. I don’t know how well it could write an entire snarky blog post.