Should AI-generated news be published without mandatory human editorial approval?
AI-generated news, when properly trained and monitored, can produce accurate, fast, and bias-free reporting at a scale impossible for human teams. Mandatory human approval creates bottlenecks that slow time-sensitive reporting, advantages platforms without ethical constraints, and does not guarantee quality — human editors introduce their own biases. Robust post-publication monitoring with AI correction systems is more scalable and ultimately more accountable than subjective pre-publication editorial judgement.
— Prof. Anand Krishnamurthy, IIT Bombay AI Ethics Lab
The proposition that AI can be trusted to publish news without human oversight fundamentally misunderstands what editorial judgement is. It is not fact-checking, which AI can do adequately. It is contextual, cultural, and consequential — it is the decision about what a story means, who it harms, what it omits, and whether it should be published at all. No current AI system has that capacity, and the cost of getting it wrong in journalism is not a software bug; it is damaged lives, inflamed conflicts, and destroyed trust.
— Sevanti Ninan, Media Critic & Founder, The Hoot