Thursday, 8.29pm
Sheffield, U.K.
A single conversation across the table with a wise man is better than ten years mere study of books. – Henry Wadsworth Longfellow
How will technologies like ChatGPT – an artificial intelligence system that can answer any question you ask – change the way we learn? What is the point of learning anything if a computer can do it for you?
Such questions must have been asked every time a new technology was invented. Writing was probably decried by people who were able to remember things in their heads. The printed book, The typewriter, the computer, voice-recognition software… Everything is changed and yet, we seem to be able to work with the changes.
For example, over the last year I have been immersed in research papers. The first section in any paper is meant to be an introduction to the area – a scholarly summary of what’s important. The problem is that these summaries are often not very good.
You see, one starts writing with good intentions – with a view to summarising the big and important ideas. But then you start to look at the text and ask questions like are there enough papers from the journal you’re writing for, should you cite some papers that the editor has written. You start to change your paper not to make it better but to get it past the person who has the power to print it. And that’s not a good way to create good material – it’s a distraction from the actual work which is to be clear about the ideas that matter.
An AI model doesn’t have those hangups. It looks at the material in its database and creates an output. I’ve tried ChatGPT in a couple of areas and in each case it’s created a coherent summary of the key ideas.
For a researcher, that’s really useful. You can have AI software that reviews all the available literature and summarises it for you in a few seconds. You don’t need to read every paper – the system does that for you.
Now, that may not seem like a good thing – shouldn’t you read everything yourself. How will you learn? But the fact is that the words in the papers don’t actually matter – it’s the ideas that do. If you can get a grasp of the key ideas, get a sense of the framework that ties them together, then you’re on your way to understanding what’s going on. The idea of a literature review is to know what’s already out there so you can build on it. But there’s so much that you couldn’t possible read it all yourself – but your AI buddy can, making sure that you don’t miss something important.
These are exciting times but the times are probably always exciting if you like learning – there’s always something new to discover.
Cheers,
Karthik Suresh
Handcrafted insight 