Being a predominantly self-interested sort, I can’t say I’m a fan of ChatGPT. Or any of its AI competitors.
Fascinated, sure. But a braying advocate? No.
I’m not naturally inclined to feed a robot that could yet do me out of a job, much less to pay it to do so. Talking of that job, it wouldn’t be right to use ChatGPT in my line of work in any case.
There’s a clear, if unspoken contract between you and me that when you read something I say I’ve written that I’ve actually written it. It’s a trust thing. (Definitely me here, btw.)
So it was with not inconsiderable reluctance that during Watches and Wonders Geneva, I logged into ChatGPT for the first time. My excuse, if there is one, is that I was moderating a panel on how AI will impact the watch industry, and in the interests of entering into the spirit of the topic, I thought I’d ask ChatGPT for an answer to that question.
Then, having got that far, I decided to ask it to provide me with some questions to ask the panellists. Can AI muster a Q&A on itself? Well, yes. And in the blink of an eye.
Did this feel right? No. No, it didn’t.
Which of course also made it rather exciting. Were the questions any good? In the main, yes. ChatGPT gave me 10, I edited them down to 8 to account for relevance and repetition, and then added one or two I felt it had missed.
We’ll come to those, but first, to its answer to the question: what will AI mean for watchmaking? In a brief summary, ChatGPT noted there would be “significant advancements”, beginning with “precision and efficiency in the manufacturing process,” followed by “quality control, predictive maintenance, and even in creating personalised customer experiences through data analysis.”
Then, rather alarmingly, it predicted this: “Overall, AI has the potential to revolutionise the watchmaking industry by streamlining processes, improving quality, and enhancing customer satisfaction.”
Big claims. Enhancing customer satisfaction? I invited panellists to pass comment, and they all seemed to agree. A CEO, a seasoned industry expert and a tech guru all nodded along with ChatGPT’s assessment and admitted they had been using AI tools for at least a year or two.
Indeed, AI is here. In fact, it’s been here for a couple of decades, but the exponential improvement curve of the past few years means the non-tech world is only now starting to really think about it.
It’s also already embedded in the watch industry. For now, we’re unlikely to hear much about it, though. Most of the use cases are invisible to the end consumer, who won’t care that Bulgari is using AI to manage supply and stock levels to make it more efficient, or that most large-scale component manufacturers are using AI to autocorrect mis-firing milling machines.
ChatGPT’s first few questions got all that out on the table. Then it grew a conscience and asked whether there were any ethical considerations in adopting AI into watchmaking, and whether any jobs would be at stake.
In the first instance, of course there are, agreed the panel. For now, AI borrows/copies/steals ideas in the absence of capacity to generate its own, and the potential for abuse of intellectual property is colossal.
Hublot chief executive Ricardo Guadalupe has twice told me his creatives have experimented with designing a Hublot watch using AI. What they got was a nasty coagulation of existing models. Panellists were also aligned that declaring the involvement of AI in the development of a watch, where it applied, would be right and necessary, too. Let’s see how that one unfolds.
Jobs meanwhile were only at risk if skilled workers were ill-equipped to use a new generation of AI tools, many of which they might already be using at home where their access to technology was less regulated. That too is a storyline to keep an eye on.
This left two questions that AI had not posed. The biggie, for me at least, was this: do watch buyers actually want watches conceived and produced by AI?
As it stands, we don’t really know. But as is so often the case, the answer when it comes is likely to be complicated. At the rarefied end of the watchmaking spectrum, where elven watchmakers craft, whittle and purr, no, almost certainly not. The human connection is where the mystery and the magic lie, and indeed where the highest margins kick in.
But at the volume end of the spectrum, will customers care? If you’ve been to Omega’s factory, you’ll have seen vast – and hugely impressive – robotic production lines where its next-gen mechanical movements come together. Does that make an Omega less attractive to its target buyer? Half a million annual customers suggest not.
We could be on the cusp of Quartz vs Mechanical Part Deux. Only this time the lines would be drawn between the small and enormously expensive volume of mechanical watches created by hand, and those generated by AI, where humans are needed only to tick boxes rather than to make watches tick.
That thought then spurred a further question: could AI actually be the saviour of the mechanical watch industry? Mechanical watches are far more expensive, all factors accounted for, than they were a decade ago. As I’ve said here many times before, volumes have plummeted and values have skyrocketed. It’s a closed shop.
What if AI could drastically reduce production, distribution and retail costs so that brands could offer Swiss Made mechanical watches in higher volume and at lower prices, opening up the industry again?
That, as a logical extension of what AI might mean for watchmaking, is surely possible. It might mean groups and brands need to create new lines, or even new brands, rather than risking the dilution of existing brand and product equity. But that’s do-able.
It also creates conflict. Because there’s limited compatibility in wanting to conserve the traditions of mechanical watchmaking while also looking to embrace technologies that might save an industry. Protect jobs and livelihoods, or reconnect with an aspirational consumer turned off by watches they can neither find nor afford? What if we want both? This tension may soon cease to be conjecture.
I can’t say I want watches designed, generated and sold by AI. Too much at stake. But then if the greater good is the preservation of a 500-year-old industry that we know and love? Now there’s a conundrum. I just hope it’s fellow humans that solve it.