Stay informed with free updates<\/p>\n
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.<\/p>\n
On the first Monday of this month, AI made it into headlines around the world more than 2,100 times.\u00a0<\/p>\n
Artificial intelligence was reported to be scamming people in Malaysia; learning the Luganda language in Uganda; reshaping the Indian tech sector, and creating billionaires across the US and China.<\/p>\n
It was a different story on the same Monday in August last year, when AI only made it into 764 headlines, according to a search of Factiva, a huge global database of newspapers, magazines and other sources.<\/p>\n
The number was just 140 in 2022, and 110 in pandemic-struck 2020. Go back a decade to 2015 and there were only nine such headlines, two of which concerned the Chinese artist Ai Weiwei.\u00a0<\/p>\n
The AI<\/a> boom that followed the 2022 release of ChatGPT shows little sign of busting. But there are many times when I wish it would.\u00a0<\/p>\n To be precise, I wish there were just one day \u2014 or even half a day \u2014 when I never read, heard or was asked about anything to do with AI.\u00a0<\/p>\n It\u2019s not so much the depressing number of times one hears that the technology is killing jobs, making us dumber, making stuff<\/a> up, supercharging fraud, stealing<\/a> creative work and driving up carbon emissions as it enriches a cluster of tech giants.<\/p>\n Nor is it the exploding valuations<\/a> of said giants, and the $100mn signing bonuses they are now reportedly offering AI research stars.<\/p>\n It\u2019s more that the rising tide of news about AI is manifestly failing to bring a commensurate level of insight about exactly how it is being used and what it will end up doing.<\/p>\n In the nearly three years since ChatGPT\u2019s launch, the gap between insider and amateur AI knowledge seems to be widening rather than narrowing.<\/p>\n Speaking as a confirmed amateur, I do not doubt the people in, say, professional services firms who claim AI-generated reports are saving thousands<\/a> of hours of work, or the researchers who say it makes it much easier to spot abnormalities on chest X-rays.<\/p>\n I can see why the AI models of Nobel<\/a> prize-winning researchers might eventually help overcome a smorgasbord of scourges, from Parkinson\u2019s<\/a> disease to drug-resistant bacteria and plastic pollution.<\/p>\n The other day I spoke to the chief executive of the Gates Foundation, Mark Suzman, who said AI models in apps are giving poor farmers real-time help to grow and sell crops. He thinks such models will have a \u201ctransformative<\/a>\u201d effect in Asia and Africa.<\/p>\n The trouble is that all this still seems so remote from so much of what we read about AI, not to mention our lived experience of it.<\/p>\n Those Factiva database searches I did to check AI headline mentions took a bit of time, and required a paid subscription. But they produced verifiable, credible results.<\/p>\n ChatGPT, on the other hand, took just seconds to misinform me that, on August 4 2025, \u201conly one\u201d headline in major news outlets explicitly included the term \u201cAI\u201d.\u00a0<\/p>\n Moments like this make me worry about the growing number of people I know who rely on AI tools to diagnose medical ailments, compose legal contracts or, most alarmingly, \u201cfact-check\u201d what they read online.<\/p>\n On the upside, my experience, slight and unscientific as it was, adds weight to a study<\/a> released last month showing that experienced software programmers \u2014 whose jobs seem seriously at risk from AI \u2014 took 19 per cent longer to do a set of tasks with AI coding tools than without them.<\/p>\n Things may change. The share of Americans now using AI daily at work is double what it was 12 months ago, even if it is just 8 per<\/a> cent.<\/p>\n But for the moment, we are stuck with mass AI confusion.<\/p>\n