{"text":[[{"start":5.75,"text":"AIs are not just machines. They are silicon wordsmiths with endless patience and flawless grammar. They can generate clean prose at scale and on demand: LinkedIn posts, scientific papers, company press releases. Let us delve into why that might — or might not — be a problem."}],[{"start":25.75,"text":"The advent of ChatGPT in November 2022 triggered much research that tried to identify “tells” in text generated by large language models. Some of them are in the first paragraph of this editorial (which was composed by a human — FT policy prohibits journalists from using AI to write): long dashes; the rhythm of three; “X with Y and Z” descriptors."}],[{"start":48.15,"text":"Barron’s recently searched for one common AI figure of speech — the “it’s not this, it’s that” phrasing used above. Its analysis of company documents, including regulatory filings and earnings statements, found “an intense ramp-up” of the distinctive structure in 2024, suggestive of widespread LLM deployment. Other studies have detected a jump in certain words in scientific papers: “underscore”, “garnered”, “intricate”, and “delve”."}],[{"start":76.3,"text":"This surge in processed prose has led to a proliferation of “AI humanisers”. They aim to turn the machine-like sheen of LLM-generated writing into something a flesh-and-blood author might have turned out, complete with flaws. Developers of some of these apps also offer the software that claims to be able to spot machine-written passages. Despite the supposed tells, chatbots have long since passed the Turing test of whether they can produce text that convincingly imitates human output."}],[{"start":105.1,"text":"Three dangers stand out. As machined “thought leadership” overwhelms our inboxes and social media, AI will start learning from itself, homogenising all prose into grey slop. Second, a race to the bottom will ultimately pit writing machines against reading machines, excluding humans. Investors already use automated programmes to dig investable nuggets from corporate press releases, for instance."}],[{"start":129.7,"text":"Then, unless we find a way to maintain and improve education in writing skills, we may confront the greatest danger. Sir Ernest Gowers hinted at it in Plain Words, his widely read 1948 pamphlet aimed at instilling a clear writing style in the British civil service: “The first requisite for any writer is to know just what meaning he wants to convey, and it is only by clothing his thoughts in words that he can think at all . . . And the less one makes a habit of thinking, the less one is able to think: the power of thinking atrophies unless it is used.” "}],[{"start":163.79999999999998,"text":"But there is no turning back. Whether it matters that we are being flooded by AI-generated communication really depends on who is doing the communicating and why. Not all text is poetry, after all."}],[{"start":176.04999999999998,"text":"Marketers have to run with the algorithm-fuelled pack, as they compete to show their worried shareholders that they can use AI to produce better campaigns, faster. AI is making some professional writers redundant but it has, in fact, improved the base level of much human-created raw copy. It is a boon to poor spellers or people uncertain of their grammar. It helps writers using a second language (particularly if it is English). It may accelerate the production of research by letting scientists concentrate on the results, without having to worry about writing the academic paper that contains them. "}],[{"start":212.75,"text":"Prose stylists must continue to cultivate the creative and human touches that distinguish their writing. But if the quality of what AIs pour on to the page pains them, they should remember that developers trained LLMs on human text. If our bots are now sending back an amplified version of the cliché-riddled drivel most of us produce, we have only ourselves to blame. "}],[{"start":239.4,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1777686596_4734.mp3"}