It seems that barely a day now passes without some new AI-related controversy appearing in the news. Particularly popular at the moment are the breathlessly reported stories from companies like Anthropic or OpenAI, etc., that reveal in hushed, conspiratorial tones how their latest AI model has managed to ‘escape’ containment and commit various misdemeanours online. Now, as marketing campaigns go, I get it: “Our AI is too clever to be controlled” is a cool hook (even if it is basically demonstrating the product failing to work as intended), but have none of these people seen Terminator? Perhaps Skynet has sent a humanoid robot back from the future to set up a guerrilla marketing agency… It’s also true that I’m not the kind of big Silicon Valley tech guru smart enough to turn US$1 of revenue into US$4 of costs, but has nobody over there considered cutting off internet access to the machines hosting these truant models?
There are also concerns about the environmental impact, primarily focused around the gargantuan resources demanded by the datacentres that keep it all running. Utah, for example, recently approved construction of a datacentre twice the size of Manhattan that will use more electricity than the entire state (up to 9 GW) and rely on water supplies already stretched by years of drought. One analysis has estimated that the waste heat generated by the facility will raise temperatures in the immediate area by as much as 2.7°C during the day and 6.6°C during the night. So much for ‘keeping 1.5 alive’… Then there’s the question of copyright, and the rather unpleasant fact revealed through court filings that at least one major AI company has been buying, scanning, and destroying books (including rare and antique editions) by the million to train its models. And, most recently, the governor of the Bank of England, Andrew Bailey, has warned G20 finance ministers that rogue or misused AI could threaten global financial stability and presents a major cybersecurity risk to financial systems, particularly as (ironically) more and more economic growth is centred on a relatively narrow band of AI companies. I have yet to work out whether Bailey is engaging in guerrilla marketing or not.
Yet, despite the feeling that we’re living through a bit of a digital Wild West, AI continues to have increasingly valuable, tangible, real-world impacts, including in the cement sector. It’s also worth noting that AI companies aren’t a monolith and that the questionable actions of some (despite making for compelling reading) don’t represent them all. AI and machine learning technologies are already helping cement producers optimise their processes in ways that were previously unimaginable. From managing the complex and ever-changing dynamics of the kiln to supporting maintenance programmes, AI is already delivering measurable improvements in efficiency, reliability, and cost. To find out more about the good that AI is doing, register free-of-charge for Optimisation 2026, World Cement’s virtual conference on 29 October, which looks at the cutting-edge technologies driving the future of cement production: worldcement.com/events/optimisation-2026
With that to look forward to, I thought it would be appropriate to end on this quote from Sarah Connor: “The unknown future rolls toward us. I face it, for the first time, with a sense of hope.”
