![\"Writer](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/06\/Writer-at-desk-e1780499599479.jpg\")
\t\t\t<\/p>\n\t\t\tThursday 04 June 2026 5:01 am
\n\t\t\t\t\t\u00a0|\u00a0\u00a0Updated:\u00a0<\/p>\n
\t\t\tWednesday 03 June 2026 4:14 pm\n\t<\/p>\n
\t\t\t<\/p>\n
City AM\u2019s inbox is drowning in AI-written pitches. Anna Moloney, who now spends her days copy and pasting them into detection software, has had enough<\/p>\n
In this slot there should have been an op-ed. A punchy one from a high-flying lawyer on the ever raging debate around inheritance tax, perfectly attuned to the interests of City AM readers. However, as is now happening at an alarming regularity, what said lawyer filed was not an insightful piece drawing on their many years of experience and expertise, but 600 words of AI-generated slop.\u00a0<\/p>\n
I will not name, but I will shame.\u00a0<\/p>\n
Unfortunately, this is no longer a rare occurrence<\/a>. In the last few weeks, running copy filed by those I thought were reputable contacts through AI detection software has become a core, and unloved, part of my job. This week alone, I have gone back to three \u2018writers\u2019 asking why their copy has been flagged as 100 per cent AI-generated with high confidence, per our software\u2019s rating. Some feign shock \u2013 \u201chow strange!\u201d \u2013 though most, I\u2019ve been surprised, fess up immediately, usually with caveats they think absolve them.\u00a0<\/p>\n They \u201cjust\u201d used it to edit down a piece to the word count. They \u201conly\u201d used it to \u201crefine\u201d a draft. They \u201cdon\u2019t know\u201d if their client used AI to write the piece, but \u201crest assured\u201d they edited it afterwards. In one case, I received a follow up, saying they were very disappointed to see the piece hadn\u2019t run, and that \u201cit would be a shame for it to go to waste\u201d.\u00a0<\/p>\n Call me selfish, but I find myself far more concerned with the waste of my own time in these scenarios. <\/p>\n At a newspaper, we run things to tight deadlines. When we\u2019ve saved a slot \u2013 such as this \u2013 for a promised piece, which then turns out to be AI-written, we are left with a hole. This is not a new problem. Writers since time immemorial have filed late, badly, or not at all. Usually in this scenario, my colleagues and I would look for someone to quickly fill the breach \u2013 as we have many talented columnists<\/a> who are able to turn around sharp copy at short notice without the assistance of AI. Needless to say, this is a skill that anyone who wants to write in newspapers, or those tasked to on others\u2019 behalf, reasonably requires. If you are unable to do that, you may need to get out of the game.<\/p>\n In my favourite defence of an AI submission, I was told that \u201cthe original thoughts and opinions of the piece\u201d were the writer\u2019s, but \u201cthe support of AI was used to help structure and organise his original thoughts into a draft format\u201d. Tell us if we\u2019re wrong, but at City AM we call that process writing. And we\u2019re actually rather keen on it.<\/p>\n AI detectors are not infallible (though I should mention we use a good one, one that seems particularly skilled in picking up copy that has been prompted to be as \u2018undetectable as AI\u2019 as possible). But there are always tells that lead us to putting it through the detector in the first place. For me, it\u2019s less the em dashes than just the banal genericism. The feeling that anyone could have written this, because anybody could have.\u00a0<\/p>\n The truth is simple, if you can\u2019t be bothered to write it, why should anyone be bothered to read it? It\u2019s insulting but, more than that, it\u2019s depressing. Writing, especially in opinion, is about individual voice: how you turn a phrase, the anecdote only you can tell, that odd word you might have picked up from a well-loved novel.\u00a0<\/p>\n If you think you can fool us, I\u2019d urge you to think longer-term, not just for us but yourself. Not only are there material costs \u2013 we are extremely unlikely to trust you enough to commission you again \u2013 but ethical ones. Trust in the media is already at an all-time low; making it a dumping ground for your Claude castoffs is unlikely to help.<\/p>\n In his now much-quoted encyclical about protecting humanity in the age of AI, Pope Leo XIV has warned<\/a> we are at risk of constructing a new Tower of Babel with the pervasive use of LLMs. At City AM, we\u2019ll put it more simply: just stop the slop \u2013 please!<\/p>\n \t\t\t\t\tRead more<\/p>\n