ChatGPT can search. It cannot research.
The platform has won the world’s attention because of the speed with which it assembles and produces information in a logical sequence, powered by artifical intelligence, as if a human being were prompting it to write. Curious about understanding what ChatGPT can do, I opened an account and asked if it would replace research in the future. Promptly, it replied: “As an AI language model, I don’t hold personal beliefs, but I can present arguments on both sides of the topic.”
The answer amused me. The speed with which it typed the information across the screen was mind-boggling. I could not read the answer as fast as it typed out the sentence. But, when it had finished, I read through the document. It was impressive. It was truly informative.
Five different points had been explained briefly in seconds. I would have thought of those five points in an hour or more, depending on the complexity of the topic. “AI: Catalyst, not Conqueror: ChatGPT and the future of research” was the title that it threw up to my question. It matched my opinion about the future of research after ChatGPT. The write-up had an introduction and a conclusion, just as our English teachers have taught us.
The document came from countless databases, as that is what it’s designed to do. By contrast, when I research a subject for an 800-word article, I visit five or six authoritative sources; I don’t go to a million as most are copies.
The difference between the write-up of ChatGPT and my research article is the former is a mere summary. When I write, I search new ideas and knit them in a way that reflect my thoughts and perspective. I choose only what is original, unique, and fresh and process them in my mind to create a fresh angle.
Research has no value if it only summarises known information. Even a literature survey in your subject requires you to say what you will do with it further, and how.
Research is not a joke. Years ago, I had asked my professor, later on my supervisor, “How can I publish in research journals?”
He answered, “When you have read everything on the subject, and have something original to add to the existing body of research.” My head swam.
Daunting task
Excellent research is a daunting task and calls for hard intellectual work. When you research, you don’t go wide; you go deep. You thoroughly investigate a minor concept, researching it extensively, substantiating each claim with reliable sources, experiments, models, and beyond. It took years to discover the atom and the theory of gravitation. ChatGPT can summarise these discoveries, but it cannot add to them. It does not have the mental wherewithal for that.
So the last question. Will ChatGPT dilute or replace seminal research? The answer is no.
A blog without sources or passing an exam through copying will not get the recognition that the work of those who conduct path-breaking research and contribute to existing knowledge gets.
The world has many unanswered questions, and will always have them. ChatGPT can put together known information but cannot ask new questions or find answers to those questions. Its limits are the limits of the team feeding the information on the platform.
ChatGPT is growing in popularity as a tool for speeding up a job and completing a project with efficiency.
Technology can be our personal genie, helping us select, classify, and analyse data we have chosen. But as researchers, it’s our responsibility to choose the focus of our inquiry.
(The writer is a former Professor, IIT-Bombay)
ceogiit@gmail.com
