Why Does ChatGPT Use ‘Delve’ a Lot?

This chatbot can answer all kinds of queries with high levels of proficiency, but the size of its vocabulary leaves a lot to be desired.

ChatGPT likes to use the same words in the same locations, and the word ‘delve’ is a prime example.

The chatbot responds to informational prompts by writing an introduction, body, and conclusion, and more often than not, the intro will contain the phrase ‘delve into’ or something like that.

Why? It’s partly to do with how OpenAI trained their AI platform to become a competent writer. Users of ChatGPT also have to shoulder some blame, because vague prompts can contribute to repetitive words and phrases too.

Pattern recognition in essay writing

ChatGPT likes to generate text using the essay format. We mentioned this already: introduction, main body, and conclusion.

Learning to write essays made sense for the development of ChatGPT for the same reasons why we get humans to write them. They foster communication skills, research skills, analytical skills and more.

ChatGPT has processed so many essays that it has even picked up on the most common essay words, of which ‘delve’ is one.

This is all well and good until a user asks ChatGPT to create different types of written communication such as articles and guides.

Cookie cutter introductions

Just like how many students explicitly state their intentions in the introduction of an essay with the phrase “In this essay, I will delve into”, ChatGPT uses similar phrasing in the introduction of AI-generated articles, such as “In this article, we’ll delve into…”

Such phrases are redundant in article writing (I didn’t explicitly state the intentions of this article in my introduction, for example), but ChatGPT doesn’t like leaving its comfort zone and so it uses the same template for everything.

After all, articles can be very informal and very personal, and it just so happens that academic writing is far from informal, and not to mention it’s very hard to express a personal opinion or display some personality when you’re not an actual person…

Bad ChatGPT prompts

The fact that ChatGPT lacks personal opinions and relies too much on a formal writing tone can be easily rectified with prompts, which in turn should put an end to the ‘delve’ problem.

If you ask ChatGPT to write an article and provide zero supplementary information about your preferred writing tone, target audience etc., then the chatbot has no choice but to do what it knows best and adopt the essay format.

Use a conversational tone as you ‘warm up’ the chatbot and feed it info about your personal opinions and experiences. This will help steer ChatGPT away from the stiff and sterile nature of essay writing and leave you with organic, human-friendly AI-generated responses.

Summary

ChatGPT commonly uses the word ‘delve’ because it is a common word in academic writing, a field in which the chatbot has received extensive training.

You can use carefully crafted prompts to deter ChatGPT from employing essay-style writing, otherwise it has no reason to change how it writes.

By giving the chatbot a personality, or straightforwardly telling it to drop the formal tone, it will become more loose and free with its choice of words and phrases.