Commissioning editor Vasudevan Mukunth captures the new existential crisis in Indian newsrooms: editors are no longer just checking copy; they are investigating authors.
AI-written or AI-polished submissions are rising exponentially. Beyond violating The Hindu’s policy, Mukunth argues they corrode the human practice of writing — the hard labour of aligning intention with precise words, wrestling with freedom a sentence affords.
His workflow is revealing. He abandoned personal “hallmarks” lists for two detector tools, GPTZero and Pangram. Independently, if both flag 5 out of 100 pieces as AI, the false positive rate is 1 in 4 combined — but when they agree on the same piece, it drops to 1 in 1,000. The problem is the middle ground: “50% chance this is AI”. Then, judgment begins.
Mukunth must balance four sensibilities: not vilify authors who use AI for non-malicious reasons, elicit honest disclosure, maintain author trust while refusing AI text, and justify reliance on tools over instinct. It is a negotiation, requiring timestamped version histories and reporting notes from rookies, and frank conversations with seniors.
His deeper insight: maths alone cannot decide. An editor must read the writer as well as the text — why did they use AI, out of insecurity, speed, or deceit? The column is thus a defence of editing as an ethical, human relationship in an era of sterilised glaze and undulating rhythm that AI offers to erase. The threat is not just fake text; it is frictionless writing.