First things first.  

No, I don’t definitively know if that viral column on the “quiet grief of adult friendship” was written using some generative AI, a lot of it or none at all. 

Yes, I’ve seen the screenshot of the AI detector, which says “100% of this text is AI generated”. 

Tools like those used by AI detector Pangram Labs claim a high degree of accuracy. But unless the columnist or publication comes forward and confirms whether and to what extent gen AI was used, we can only make an educated guess – and so far, neither have. Both parties may have little incentive to do so, too – the column has probably racked up lakhs of views for the platform, and the reputation of the columnist, who is in public services, is at stake. In the absence of that clarification, one can only say that some of the phrasing, the structure and the cadence seem similar to writing produced by Chat GPT. 

Incidentally, the same columnist, whose bio says he is an Indian Police Services officer, has written another 14 columns for the Times of India website in the last 11 months, seven columns for India Today in the last five and a few for The Print. Some of the ones I skimmed through had a similar style and phrasing  – but again, it’s hard to establish beyond doubt whether or how much gen AI may have been used.

For the sake of this column, though, it doesn’t matter. I am not here to incite an online witch-hunt. I was more interested in the reactions to the article. At a time when our algorithms are shaped by manufactured virality, paid influencers and bot accounts, this blew up the old-fashioned way – organically. Blue-tick and other users on X posted about how beautifully it was written and how much it resonated with them. The article stepped out of the Twitter bubble and coursed through WhatsApp, with messages discussing how everything it said was so true, so relatable. Screenshots were highlighted and shared, sentences were parsed and passed around. Some talked about the friends they held close, others about how they would reach out to old mates. It struck a chord.

All of that was real and human. Which now may or may not be ironical.

I understand the feeling of disappointment on discovering that AI may have been “secretly” used – it’s like being conned. And one can legitimately take issue with someone passing off an anecdote and emotions generated by AI as their own – that’s dishonest, especially in a column. But that apart, how much should it matter whether it was written entirely by a human? If the writer was “assisted” by AI, even extensively, does it take away from how the article made you feel, and the appreciation for the writing? What if the article came with a label clarifying that AI was used in its composition – would that change how you read the piece or connected with it? Does liking something written using AI make one inferior in any way - is it a moral judgement? 

There are some on X who viewed what had transpired as a lack of discernment about good writing, or a consequence of people not reading enough. It may be a consequence of people not reading enough AI writing, perhaps, making it hard for them to identify the “tells”. But good writing can be somewhat subjective – someone’s mediocre writing might be someone else’s excellent writing. Who’s to decide and why should they listen to you? What is indisputable is that the column appealed to large sections of people who felt they identified with what it expressed and enjoyed the writing, literary criticism be damned. 

If all of this sounds like I’m shilling for AI-generated columns, writing or journalism, that’s not my point – that will put me out of work, for one. But the truth is, we are living in a world where we are going to see a lot more of this. And it’s not just regular readers (if there is such a thing) who are being bamboozled – the prestigious Commonwealth Short Story Prize is currently roiled over charges on social media that one of its prize winning short stories was, again, generated by Chat GPT. The jury commended the writing for its “lyrical precision..haunting atmosphere.. Confidence and restraint of its voice”. People on X called it bad. On its website, the Commonwealth Foundation now says it is conducting “a thorough, transparent review of the selection process” and that all entrants are asked to submit “their original work”. 

As the models get better, it’s possibly going to get more difficult to make out whether AI was used in the writing process. We are also likely to see more readers suspicious of whether something has been written using Chat GPT and its ilk, and pounce on it out on social media – which could go either way.  

Ultimately, publishers, platforms and media organisations need to take a call on how they are regulating the use of AI by its employees and contributors and figure out how to enforce their stance and standards, whether it’s banning it entirely or publishing with a disclosure. But when an article goes viral, will some of them even care? As readers, the question we may increasingly face is, would we? 

The response of one reader on X who shared the column with the comment that it was beautifully written was telling. When it was pointed out to him that 100 percent of the text was AI generated, his reply was that he didn’t care much. His take: “I dare say most things one reads now are either partially or substantially written by AI. If the work is resonant, its authorship is only slightly of interest to me.”

Will more people start subscribing to this view? Will we reach a point when we see books “co-written with AI” in fine print flying off the shelves, despite the label? Will “hand-written by human” command a premium – monetarily and otherwise? 

This last is something I would obviously hope for. As for the others – we’ll find out soon enough. 

We take no ads, bow to no government or corporation, and answer only to you, the reader. This Press Freedom Month, pay to keep news free.