A growing concern has emerged within the public relations and journalism sectors as AI detection tools are increasingly flagging human-written content as machine-generated. Emma Divers, a PR professional with a decade of experience, has highlighted the absurdity of being told by AI detectors that her own work, crafted through years of training, appears to be written by an algorithm.
Divers notes that the very techniques she was taught to employ for clear, engaging communication – such as short sentences, clear structure, minimal jargon, and the effective use of a 'rule of three' – are now precisely what AI detectors identify as indicators of artificial intelligence. This has led to a bizarre situation where she and her colleagues are compelled to deliberately introduce stylistic imperfections, awkward phrasing, and grammatical quirks into their writing to ensure it scores as 'human' rather than 'AI polished'.
The impact of this trend is significant, with one colleague reportedly redrafting a deeply personal and thoughtful opinion piece four times to satisfy an AI detector. The original, which was well-structured and emotionally resonant, was penalised for its clarity and logical flow, forcing the author to break up paragraphs and remove impactful lines to appear more 'human'. Divers questions the efficacy and fairness of AI-powered detection, especially when some of these same tools offer a 'humanise' button, effectively providing a workaround for those trying to evade detection.
While acknowledging the legitimate concerns of publications striving to maintain standards against a deluge of poorly generated AI content, Divers argues that the current approach risks becoming a 'witch hunt'. She suggests that the focus should instead be on the core qualities of good writing: storytelling, unique insights, personal perspective, and the ability to provoke thought – elements that AI, despite its advancements, still struggles to consistently replicate. The fear is that an over-reliance on automated detection could inadvertently penalise effective, human-crafted communication.
This situation presents a paradox where clarity and conciseness, once hallmarks of good writing, are now viewed with suspicion by algorithms. The broader implication is a potential shift in writing styles, where authors might feel pressured to sacrifice quality and impact to bypass detection, leading to a decline in the overall standard of published content if the current trajectory continues unchecked.