ChatGPT—interesting responses: not so terrifying yet?

Dear Editor-in-Chief:

Your editorial [1] was certainly timely and interesting. In just a few months, ChatGPT (within a few days of GPT 4 launch on March 14, 2023) seems to have taken the world by storm. A former trainee and now a junior consultant was suggesting ChatGPT to me. He wanted me to try it out. Your editorial a couple of days later was informative and I decided to try it out. One of the concerns in the various discussions was that it might replace a doctor and may suggest diagnoses, investigations, and treatments.

I got the following response to my question:

Q: What safety do you ensure for medical consultation of a patient who does not want to consult a physician?

ChatGPT: “As an AI language model, I do not have the ability to ensure safety for medical consultation of a patient who does not want to consult a physician. However, I can provide some general information that may be helpful.

It is important to understand that medical consultation with a qualified physician is usually the safest and most effective way to address health concerns. Physicians are trained to evaluate symptoms, make diagnoses, and develop treatment plans based on the latest medical research and clinical guidelines. They also have access to a range of diagnostic tests and treatments that can help identify and manage health issues.

If a patient does not want to consult a physician, there are some potential risks involved”

The response was fast but long winded. It clearly was playing safe and did not appear to act like a qualified medical professional!

I wondered if ChatGPT could really be of help to reviewers/Editors and perhaps help speed up the process. The responses to my next questions also are interesting, but don’t really seem to help. However, the responses were quite fast—barely taking a few seconds! Sample the following:

Q: Can you highlight plagiarism in this article - Indian Journal of Thoracic and Cardiovascular Surgery (May–June 2023) 39(3):217–221?

ChatGPT: “As an AI language model, I do not have access to any specific articles or journals, as I do not have the capability to browse the internet. However, I can give you a general understanding of what plagiarism is and some common examples of it.

Plagiarism is the act of using someone else’s work or ideas without giving them proper credit or attribution. It can take various forms, including copying and pasting text from a source without citation, paraphrasing someone else’s ideas without giving credit, or submitting someone else’s work as your own.

In academic writing, plagiarism is a serious offense and can have severe consequences, such as losing academic credibility and damaging professional reputation. Therefore, it is essential to give proper credit to the original sources and avoid plagiarism by using appropriate citation styles and referencing.

If you suspect that there may be plagiarism in an article, you can use plagiarism checker tools to check for similarities between the article and other sources. These tools can help identify any potential cases of plagiarism and ensure that proper citation and attribution are given.”

Q: I am not satisfied with your response, because I have sent the link to the article can you not check the article? You are known to be a plagiarism checker. Your developers are thinking of using a watermark to highlight plagiarised sections in any article

ChatGPT: “I apologize for the confusion. As an AI language model, I do not have the capability to browse the internet or access external links. Therefore, I cannot check the article you provided for potential instances of plagiarism. However, I can tell you that plagiarism can occur in various forms, including the use of someone else’s work without proper citation, paraphrasing someone else’s ideas without giving credit, or submitting someone else’s work as your own.

To detect plagiarism in an article, you can use plagiarism checker tools that compare the text in the article to other sources to identify any similarities. Some plagiarism checker tools also highlight potentially plagiarized sections in the text or provide a percentage of similarity to other sources.

Regarding the use of watermarks to highlight plagiarized sections, that is a technique that some plagiarism checker tools use to indicate potential plagiarism in a document. The watermarks can help identify specific sections of the text that may have been copied from other sources and make it easier for the user to review and correct any potential plagiarism.”

Q: Isn’t plagiarism checker tool also based on machine learning?

ChatGPT

“Yes, plagiarism checker tools use machine learning algorithms to compare the text in an article to other sources to identify potential instances of plagiarism. These algorithms are designed to analyze the text, identify patterns, and compare them to other sources to determine if there are any similarities.

The algorithms used in plagiarism checker tools are trained on large datasets of text to identify patterns and similarities between the text. The more data the algorithm is trained on, the more accurate it becomes in identifying potential plagiarism.

Machine learning algorithms can also be trained to highlight specific sections of text that are potentially plagiarized. Some plagiarism checker tools use a color-coded system to highlight text that may be plagiarized, making it easier for the user to review and correct any potential plagiarism.

Overall, plagiarism checker tools use machine learning algorithms to compare text, identify patterns, and highlight potential instances of plagiarism in an article.”

Q: Then how can you help a reviewer of a peer reviewed Journal?

ChatGPT: “As an AI language model, I can provide general guidance and assistance to reviewers of peer-reviewed journals. Here are some ways I can assist:

1. Identifying potential instances of plagiarism: Reviewers can use plagiarism checker tools to identify any potential instances of plagiarism in the submitted article. As a language model, I can provide guidance on how to use these tools and suggest some commonly used plagiarism checkers.

2. Reviewing the language and clarity of the article: I can help reviewers identify any issues with the language and clarity of the article. For example, I can provide suggestions for rephrasing sentences or using more appropriate vocabulary.

3. Providing background information on specific topics: If the article covers a topic that the reviewer is not familiar with, I can provide background information on the topic to help the reviewer better understand the content.

4. Suggesting improvements to the structure and flow of the article: I can provide feedback on the structure and flow of the article and suggest improvements to make it more organized and easier to follow.

Overall, as an AI language model, I can provide assistance to reviewers of peer-reviewed journals by offering guidance on plagiarism checkers, reviewing the language and clarity of the article, providing background information on specific topics, and suggesting improvements to the structure and flow of the article.”

Q:Thank you but I am not impressed. The language editor of Microsoft already is there, to help. Can you comment on a piece of text from an article if I send it to you?

ChatGPT: “Yes, I can definitely help you with specific pieces of text from an article. As an AI language model, I can analyze the text and provide feedback on grammar, syntax, and other language-related issues. Please feel free to send me the text, and I will do my best to assist you.”

“Thank you , I will do it at a later date.”

ChatGPT: “You’re welcome. Please feel free to reach out to me whenever you need assistance. I’ll be happy to help you.”

Sure

 ChatGPT: “Great! If you have any other questions or need any assistance in the future, feel free to ask. Have a good day!”

I think, we have miles to go before any AI can faithfully replace the human mind. We, as medical professionals, as ChatGPT has shown in its response, need not worry about being rendered expendable any time soon.

However, all technology to replace the human mind may lead to imperceptible changes in the brain that may take several generations to manifest as smaller sized brains and decreased memory, decreased density of brain dendritic circuitry, and white fibre connectivity of the growing brain. Considerable research in neuroanatomy and evolution of the brain using modern magnetic resonance imaging techniques and studies across species including fossils appears to suggest that at birth there is little difference in the size and shape of the brain of a human neonate and a neanderthal neonate [2–4]. After birth, there is a sudden spurt in growth of the human brain up to 2 years that is absent in the Neanderthal neonate. The brain continues to grow into adulthood. This increase in brain size largely reflects the increased circuitry showing increased dendritic density. Disuse/underuse atrophy is seen in many tissues of the body. The brain perhaps is no exception. The brains of only the developers of AI will probably be unaffected! When mobile phones with storage of long lists of contacts became ubiquitous, it was clear to us all that we have lost the ability to remember even the most frequently used phone numbers. When calculators appeared everywhere, they were welcomed for everyday use and in complex scientific calculations. But the three R’s continue to be taught in schools. With computers and mobiles dominating, handwriting is fast becoming a lost art except for those who use some note writing apps or phones. When tusks became a source of money to finance the wars in Mozambique, over about 15 years, the population of tuskless female elephants increased from about 2–6% naturally to 33% post poaching and post wars. This is probably a natural selection in play, as a means of survival (https://www.google.com/search?client=firefox-b-d&q=YouTube+Tuskless+female+Elephants+mozambique#fpstate=ive&vld=cid:4387604d.vid:IxJDUrDH9v4).

Let us hope that we continue to have the ability to discriminate and use any technology that does not adversely impact the development of the brain and the finer elements of emotions, empathy, aesthetics, art, and culture—that is the greatest piece of evolution that distinguishes us as Homo sapiens.

Funding

None.

Declarations

Ethical approval

Not applicable.

Informed consent

Not applicable.

Conflict of interest

None.