How ChatGPT reveals itself in one word: experts have found a sign of AI content
In the modern world, the creation of text materials by artificial intelligence has become commonplace. However, upon closer inspection, they can be easily identified by the use of certain words, which is noticed by experienced specialists.
ChatGPT indicator word
Unusual vocabulary
Paul Graham, co-founder of the famous American business incubator Y Combinator, discovered that inclusion the words “delve” (to go deeper, to dive) into the body of the email is a sign that it was created by ChatGPT. It is a relatively rare word that appears more often in academic works and is used to give speech an appearance of intellectuality.
Usage statistics
Data show a sharp increase in the frequency of use of this word in scientific materials: from less than 2,000 occurrences in 2022 to almost 18,000 in 2024. This trend is explained by the widespread use of ChatGPT and other text generation systems based on artificial intelligence.
Popularity with ChatGPT
The AI Phrase Finder service, which analyzed 50,000 ChatGPT responses, also revealed that “delve” is one of the ten most frequently used words by this AI. Graham noted that people resort to such expressions in an effort to add an intellectual touch to their speech.
Discussion and criticism
Many commentators, especially from the former British colonies, perceived this conclusion as discrimination against certain linguistic groups for whom "delve" is in an ordinary word. However, Graham countered that children today learn English primarily from films and other media, rather than from books.
Glossary
- Y Combinator - a famous American startup accelerator
- ChatGPT - a popular text generation system on based on artificial intelligence
- AI Phrase Finder - a service for analyzing texts created by AI
Links
Answers to questions
What are the main characteristics of the product?
What tasks is this laptop suitable for?
What are the advantages of this laptop over its competitors?
What features make this laptop a better choice?
Is this laptop suitable for work and play?
Hashtags
Save a link to this article
Discussion of the topic – How ChatGPT reveals itself in one word: experts have found a sign of AI content
Y Combinator co-founder Paul Graham discovered that the presence of the word “delve” in an email could indicate that it was written by ChatGPT. This word is often used by AI to demonstrate deep immersion in a topic.
Latest comments
15 comments
Write a comment
Your email address will not be published. Required fields are checked *
Иван
Interestingly, even words like “delve” can give away the spelling of ChatGPT material 🤔 People usually use simpler words in conversation and correspondence.
Анна
Yes, Ivan, I agree. Not bad noted 👍 Although words like “delve” may well be found in an academic environment or among people with a large vocabulary. You can't judge by just one word.
Марио
Guys, don't you think this is a little discriminatory? 🤨 People from former British colonies may well use such words in everyday speech. This is normal for them.
Эмили
Exactly, Mario! 👏 You shouldn't make hasty conclusions. Vocabulary depends on many factors: education, origin, hobbies... Judging by only one word is a simplification.
Курт
Hmm, I recently did a survey and it turned out that most of my friends don’t even know what “delve” means 🧐 So for me personally, this is a rather unusual word in regular correspondence.
Ольга
Yeah, it’s a controversial issue 😅 On the one hand, the use of such words can give out ChatGPT. On the other hand, you cannot judge people by their vocabulary. A more subtle approach is needed.
Франц
Recently I came across an advertisement for a new service that can identify texts written by ChatGPT 🤖 It analyzes not only words, but also style, sentence structure, etc. Maybe this is a more reliable way?
Джулия
Interesting topic! 🤓 Did you know that the word “delve” comes from the Old English “delfan” and meant “to dig, to dig”? So it may well be used in academic circles or by people with good knowledge of the language.
Збышек
Brrr, come on 😒 Why even bother trying to distinguish ChatGPT from a human? Technology is moving forward, nothing can be done about it. It is better to immediately accept this as a given.
Анна
Zbyshek, you are right that technology is developing 👍 But it is still important to distinguish a person from an AI in some situations. For example, when applying for a job or university. Otherwise you won't get away with fraud! 🙃
Марио
Wow, what a twist! 😮 I didn’t think that the word “delve” could cause such heated debate. At the end of the day, the main thing is not to judge people by their vocabulary, but to respect diversity and cultural differences.
Франц
By the way, did you know that in the latest versions of ChatGPT the developers specifically removed some unusual words like “delve” to make identification more difficult? 🤯 So now this criterion alone is unlikely to work.
Эмили
Oh, technology does not stand still! 🌐 Franz, thank you for the interesting information. In any case, it seems to me that the best way is a comprehensive text analysis based on many parameters. This is the only way to more or less reliably distinguish a person from an AI.
Курт
Hah, well then it’s time for me to update my vocabulary so that I don’t get confused with ChatGPT 😂 Maybe I’ll start inserting “delve” into every other sentence?
Джулия
Kurt, maybe it's better to just be yourself? 🙃 In the end, the main thing is to be humane and authentic in communication. Let technology develop, but people will always be unique!