Since AI first reared its ugly head, I’ve been keenly observing its progress as a ski review creator (https://realskiers.com/revelations/chatgpt-ai-has-ski-patter-down-cold/) with a sense of impending doom. My second peek at its progress, The Con is On!, was one of the most read Revelations of last season, an indication that many of my Dear Readers share my curiosity about AI’s infiltration of our little world.
Now that AI has had another season to sharpen its skills, I thought it might be illuminating to check in on its evolution. Here’s what I discovered.
While AI remains a formidable mimic, and gets enough facts right to dupe those with little prior knowledge of ski attributes, it has several glaring tells that tip its hand. Its biggest shortcoming is that its bottomless appetite for data hasn’t quite caught up with the present day. When asked to compose a 300-word review of the 2025 Blizzard Anomaly 94, it didn’t flinch, but nor did it have a clue what characteristics the ski did or didn’t possess. It wasn’t even sure if the Anomaly 94 was a ski or snowboard.
Not that this deficiency slowed it down. It still whipped out a review in less time than it will take you to digest this sentence. It peppered its piece with the usual hyperbolic claims, like “revolutionary design,” “advanced damping technology,” and “lightweight yet sturdy.” AI’s specialty is evasive twaddle that sounds meaningful without depending on any actual facts: “hybrid camber profile,” “premium materials,” and “advanced damping technology” all pretend to mean something but are actually information-free.
When prompted to create a 300-word review as if I had written it, AI fumbled badly, writing the usual tripe about a mythical “XYZ” model, an open admission that it doesn’t need to know the model name – or anything else of a remotely factual nature – about a ski to write expertly about it. The review was such a vaporous botch that I decided to give AI another swing at the pinata, so I asked for another, Realskiers-style review. (If you’d like to hear what AI wrote for me this week in its entirety, I’ve recited them all in this week’s podcast.)
This time AI swung even more wildly, digging up a hypothetical snapshot of the 2022 Atomic Redster S9 FIS GS [sic], its name a heady blend of both Atomic’s Slalom and GS Race skis. Even if such a ski existed, I wouldn’t have reviewed it/them, as I’d (regrettably) dropped Non-FIS Race skis from coverage some years prior. Hard for AI to craftily imitate what it can’t find and/or never existed.
But AI has an admirable “can-do” attitude about it. Undaunted by its stumble out of the gate, the software rallied and extolled the virtues of what sounds like one terrific ski: “The Redster S9 FIS GS is a true powerhouse on the slopes, thanks to its Titanium Powered laminate construction that delivers a smooth and stable ride at high speeds. The ski’s Servotec technology enhances agility and control by dynamically adapting to the skier’s input, allowing for effortless turns and precise handling through variable snow conditions.” It reads like a perfect recipe for Word Salad: take two capitalized features, one inflated claim and a soupcon of benefits of fuzzy prominence; toss lightly without irony, and serve.
Oh, and good luck with the variable snow conditions on an Atomic race ski.
The AI bot I worked with was on more confident footing when contriving a review of a mainstream ski of its own choosing, as if written by a digital magazine writer/editor. It chose the Blizzard Bonafide, giving it an enormous archive to digest and regurgitate, given the Bonnie’s longevity and sustained popularity. The review would have helped its credibility if it hadn’t called the Bonafide “all-new,” but let’s not quibble. The software wanders away from its strong suit with such babble as, “the Blizzard Bonafide sports a sleek and modern look that is sure to turn heads on the slopes.” And its prosaic prose vacillates between hyperbolic panting and breathless awe. If this sounds like some of the dreck you read online, it’s because it was born in the primordial soup of the thousands of reviews it has absorbed into its DNA.
To be fair, AI lends itself to editing and re-writes, so some errors in its first drafts are readily righted. While it doesn’t hesitate to compose a review about a ski it knows nothing about, these more fanciful efforts (like the one on ski XYZ) can also be tweaked to create a more plausible product. One glaring omission in any AI-generated review is the absence of quotes from humans, perhaps a rare instance of concern about plagiarism and AI’s unblushing theft of other people’s work. Since the truth is never of much interest to AI, I don’t see why it would hesitate to create quotes from whole cloth.
The real danger of AI doesn’t lie in the larcenies and follies it commits on a daily basis, but in its inexorably expanding knowledge and sophistication that no one seems willing or able to check. Fabricating misleading ski reviews strikes me as the least of our worries.
Related Articles
The Road to Perdition
The road to hell is said to be paved with good intentions. In my experience, the friends and relatives of prospective boot buyers are a wellspring of wretched advice wrapped in bright ribbons of sincerity and concern.
(Let us pause a moment and prayerfully acknowledge the gratitude of bootfitters everywhere that the new, pandemic-driven bootfit protocol discourages the presence of a bootfit entourage composed of family, moral supporters and consiglieri.)
Back to the subject at hand, the particular nugget of advice I’m leery of is the customary admonition to avoid too stiff a boot as it will hurt, you’ll hate it eventually if you don’t detest it immediately, and it will inhibit your skills development. Get only as much boot as you need and no more, goes the conventional wisdom. Racers need stiff boots; you don’t.
The Making of a Skier, Part IX: The ASTM, Carl Ettlinger and I
One of the many hats I wore as North American binding product manager for Salomon in the early 1980’s was that of delegate to the American Society for Testing and Materials (ASTM). I believe the first meeting of F8.14 – the sub-committee on ski safety – that I attended was in Pennsylvania. I was flying under the wings of Salomon’s seer of all standards and patents, Gilbert Delouche, and the binding product manager for the North American zone at that time (and my mentor), Joe Campisi.
I was a babe in the woods, but I soon caught on to the game under Delouche’s patience guidance. I recall a debate on the binding specification then being batted around in the technical committee chaired by Carl Ettlinger. Ettlinger wanted language that would require any release/retention setting of 10 or above to be “visually distinctive” from the rest of the scale.
In Memorium, Carl Ettlinger
Carl was a giant of a man whose outsized voice roiled every conversation like a burst dam and whose expansive vision reached across the mixed milieus of research, journalism, risk management and education. I knew him when he was at the peak of his powers, as he explained to me when I interviewed him for a “where are they now?” profile in Skiing History. He was able to conduct long-term research on injury patterns as well as analyze the particulars of the current binding market, turn around and package this knowledge into articles for Skiing and Skiing Trade News, followed up by a workshop tour that would bring enlightenment to the grassroots level. No one but Carl could have pulled this off, and Lord knows no one has had the requisite talent, energy and will power since.
But time and tide wait for no man, and Carl’s finely spun web of influence was eventually plucked apart. The loss of his pivotal positions in the press allowed him to slip from public view before we, the skiers of the world, realized we hadn’t taken the time to thank him.
We have the time to thank him now.
So thanks, Carl, for being first and foremost a teacher, for teaching is at the heart of the evangel’s mission.
Thanks for being so damn stubborn. Your insistence on improving skier safety wore through a wall of resistance as tough as Vermont marble.
Thanks for having a heart as big as that melon-sized head of yours. The fuel to your tireless mind was a caring heart that tried to embrace the world.
Thanks for all the stories once the Mount Gay flowed. Who knew we would have won the Vietnam War if only his superiors had listened? I can’t remember exactly how – he wasn’t the only one drinking Mount Gay – but I recall the light in his eyes as he relayed his twisted tales, taking us down successive rabbit-holes of digression that I lost track of at the seventh level.
That’s what I remember most vividly about my many interactions with Carl: his brain so teemed with thoughts he rushed to get them out in a verbal jailbreak that would travel around the cosmos until returning, many lost minutes later, to the subject that had inspired them. That was Carl: too many words for one sentence, too many tasks to tend to and all of it, every erg of his endless energy, devoted to a cause he never ceased to serve.
Fare thee well, Carl Ettlinger. The world misses you already for it will never see another quite like you, whose every moment seemed larger than life itself.
I raise my glass to you, old friend. Mount Gay, of course.
Jackson Hogen
June 23, 2020
