ChatGPT and the future of writing
Whether Artificial Intelligence replaces authors, teachers and artists entirely or just helps them be more creative and productive, it’s certain that it will transform the way we read, write and learn
When I first started researching this piece, I was, I must admit, left a bit speechless – sweaty palms and cold feet. Started by a San Francisco based company, Open AI (Artificial Intelligence) has a few features on its website that will forever change how we read, write, analyse, learn – or educate our children. The powerful artificial intelligence tool from this foundation funded by Elon Musk can write original content on just about anything and could, in a few years, leave professors, programmers and small-time writers jobless.
Grudgingly, I opened the surprisingly simple interface on ChatGPT – a neat box where I could type in a question or a statement and read as the AI spits out answers. ChatGPT derives its content from the internet. How then is it different from Google, you ask. While Google responds to a query by fetching relevant information in the form of many web pages, ChatGPT presents large chunks of information that is precise and all-encompassing. Any simple text like – “Can you write a poem on masculinism in the style of Chimanda Ngozi Adichie?” will give you human-like answers that are “original” and “largely accurate”.
Available for free since November last year, it can write poetry, essays, stories, movie scripts and even generate website code snippets. The website’s Dall-E feature helps generate original images or paintings according to your commands.
The OPEN AI website was just the beginning. I grew progressively wonderstruck as I learnt that authors have been using AI to write books even before ChatGPT came in. Ted Chiang’s The Lifecycle of Software Objects and John Lanchester’s The Insight are prominent examples of novels written using AI. Postmodernism Generator is also being used by school students to generate fake and largely absurd scholarly writing. Jennifer Lepp, who self publishes mystery books on Kindle ensures she is prolific by using the popular AI tool Sudowrite. On it, she outlines a scene, presses the expand button and lets the program do the writing. After making edits on the AI-produced content, she pastes it back and presses “continue”. Whenever the AI did not go the way she wanted it to, the author would start writing her own sentences again before letting it continue. Within a few days, Lepp’s productivity had increased by 23 percent. Her readers could hardly make out the difference.
While I am only just discovering AI in writing, back in late 2021, Joanna Penn, author of best selling thrillers started online classes teaching AI-writing, editing and plotting. She also used the DALL-E function to generate book cover designs. Penn approached Orna Ross, a historical fiction writer and the founder of the Alliance of Independent Authors, a UK professional organization, and together they collected feedback from peers to formulate a code of ethical AI conduct. The code states that “humans remain responsible agents” and anything produced by AI shouldn’t be discriminatory or slanderous. It shouldn’t be a shoddy “cut and paste” job and the use of AI should be disclosed to readers “where appropriate”, basically leaving an individual to make their own ethical decisions.
So should all of us be very afraid? ChatGPT can definitely produce its own stories, movie scripts and novellas, and they are quite impressive. But for the most part, it still seems like a circus acrobat performing the big top’s most popular old tricks – a collective reflection of the content available on the World Wide Web.
Movie makers and series developers have been using popular data and reviews to build generic middlebrow content for years now; but eventually the audience and readership want something original. What then?
Everywhere I turn, everyone from journal producers to writers want to use the AI on peripherals. While some use it to edit grammatical errors, others use it as a thesaurus and still others turn to it to write blurbs. I too see it as a resource finder or as something that will help me explore fresh dimensions of a book that I might be working on.
When I entered something as simple as “Books on tribals in India”, it came up with options such as Tribal Identity and the Modern World: The Changing Context of Indigenous Politics by Jens Lerche, an important book that I missed reading while conducting the research for my last book, White as Milk and Rice. As I made my questions more acute, the answers kept getting better. An author could use it to draw parallels or develop a “yardstick” essay. However, the single most important concern is plagiarism, which could be impossible to prove since the AI is meant to sound “human”.
The technology isn’t there yet but it’s certain that even the most traditional writers will have to accept AI. Remember when Google was just a search engine? Soon it became our letterbox, maps, office diary, calendar and so much more. AI could become all-encompassing in similar ways. But also, a large part of understanding AI is that it can never replace human writing, because writing is thinking. AI may be intelligent, but it cannot think. Yet.
Nidhi Dugar Kundalia’s latest book is White as Milk and Rice- Stories of India’s Isolated Tribes
The views expressed are personal