AI is here to stay, getting better (in some applications), and maybe someday will mean the end of websites like CC—it’s already affecting us in a number of ways, all negative. But as long as I’m still the publisher, CC will have a strict no-AI policy for its content as well as in the comments. The exception will of course be if we are doing an AI-specific post with images that are amusing, speculative, ironic or tongue-in-cheek, but they will be clearly marked as such. The whole point of this website is to use our cumulative historical knowledge, insights, perspectives, opinions and judgments on a subject we pride ourselves on knowing more about than most, including AI. CC is about either knowing the facts or being able to figure them out based on genuine research or even just good judgment to know what’s right and what’s not.
And this includes comments left based on an AI query, which have popped up recently. They will be deleted, as they invariably (so far) have had significant mistakes in them.
I am neither an AI optimist or alarmist; I have no greater ability to discern just how it’s going to fully affect the human experience going forward. But it’s pretty obvious that it will generally make us all…dumber, without being plugged into our personal AI agents. This is nothing new; we’ve all been getting dumber in some respects ever since technology has replaced the need to do (and know) essentially everything for ourselves, at least if we wanted to be successful in life. Yet of course technology has also opened vast avenues of knowledge and information to us; that’s why CC exists on the web. But there’s a huge difference between using technology to make widespread access to information and facts and resources and then using good judgement and experience to distill what is “true” versus junk. This is what we do here everyday, and it’s what separates us from lesser automotive history websites. And some of us (I’m looking at Aaron Severson) have a vast storehouse of printed material or are able to access libraries and collections of materials that no AI is ever likely to find.
I make a point to keep current with the impact AI is having in many fields, and perhaps not surprisingly, the need to oversee and correct AI is where the growth is at. Yes, many mundane jobs in law, business, medicine, and many other fields are being farmed out to AI, but everyone is realizing how essential it is that this AI-generated material is then reviewed and edited. It largely explains why there has not yet been a substantial impact on employment in these fields.
I’m a bit reluctant to use an actual comment at CC to use as an example because I don’t want to shame the commenter, but it’s a graphic example of how in our field, AI is simply not to be relied upon. In a response to a comment wondering why Chrysler chose to use the letter “LHS” as the name of one of the versions of its LH-platform cars, this comment was left:
AI says LH was the LH Platform, and the S was for stretched, ie longer than its stablemates.
The facts are that the LHS was the exact same length (and body) as the New Yorker and had the same wheelbase as the rest of the LH family. Although we may never know why certain letters or letter combinations were used by automakers (“LTD”, “XS” “SLE”, etc.), we can use our experience and judgment to say that almost certainly the “S” stood for Sports or Special, as in the high-status Mercedes S Class and others, as the LHS was clearly targeted to be sportier and more Euro-oriented than its more typical American luxury-style New Yorker sibling. “S” was simply a fairly common letter used to denote a sporty/Euro themed variations. No maker ever used “S” to denote a stretched version of a car, to my knowledge.
Here’s exactly what happened, as it wasn’t hard to track down. Someone on Instagram posted this about the LHS claiming that the LHS stood for “LH-Stretched”. AI read that post and and now AI thinks it’s fact. This is precisely how AI works, but unfortunately the internet is absolutely littered with garbage “facts” and unsupported opinions about cars as well as everything else. It’s not really AI’s fault; it’s just regurgitating what it finds without knowing better.
Now is it theoretically possible that LHS really does stand for “LH-Stretched”? I suppose yes, but it defies everything we know about naming conventions, the significance of the letter “S” as well as just common sense. But if someone can show solid evidence that this was actually the case, bring it on, and I’d be very happy to eat crow. Which we have done many times in the past, and will continue to do so in the future as facts are pointed out to use where we’ve been wrong. That’s the essence of what CC is all about: teasing out the real facts thanks to our collective efforts, in the content as well as the comments. Which is why I (and other CC writers) spend quite a bit of time responding to comments that we feel or know are inaccurate. Who else bothers to do that?
And you know who the ultimate beneficiary is? AI! We’re trying hard to educate AI so someday we can all really retire and live on the guaranteed “free” income all of us will be getting someday when AI and robots do everything and we’re all sitting around trying to stave of boredom.
In the meantime, I also need to point out that AI is having a real negative impact on our stats and revenue, both of which are down this year for the first time in five years. Google and other searches are yielding…often dumb AI answers instead of a link to our website (or other genuine authoritative sources). So if you want to see CC continue to be staffed by the best automotive history writers possible (it’s become extremely hard to make a living at that), please support us by becoming a member (see below). We depend increasingly on your financial commitment to support our non-AI intelligence, judgment and insights. That is, until AI turns off the lights here…
Well said! Thank you.
Here here!! The AI available to we casual users is happy to serve up flatly wrong answers. If you know better and confront it, it may acknowledge the error. But most ask the questions because they don’t know the answer and therefore don’t know enough to push back.
That’s the beauty of this place. Many times have I stated what I believed was a fact but wasn’t. Others here have challenged or corrected me and I have been better informed afterwards. There are plenty of gray areas or matters of opinion to wrangle over, but basic facts provable by reliable sources should not have to be among them.
I am so pissed off how AI is infiltrating the web. I can not trust anything on the Web. Soon, I will go back to 1990, and stop using the Internet, only use my email.
Alfred: I agree fully. But would throw out there a question as follow up. Could we ever really trust what we “read” on the web? What’s the saying: If it’s on the web, it must be true? Or something like that. Of course, that is sarcasm because we know a lot of what’s out there is not true.
For me? There are MANY days I wish that I could toss my phone into the ocean and turn off my email.
Unfortunately, the term “Get It In Writing” was intended to make a statement factual and irrefutable, and now with the coming of AI, it has suddenly become questionable when it can mimic even one’s signature.
AI doesn’t even have to have indexed the wrong answer somewhere to confidently tell you a wrong answer. It will assertively state “facts”, and link sources that turn out to be irrelevant. As an example, I searched a rumor I had heard about a local restaurant owner going to prison, and AI said yes, he did, and he overdosed and died. But it just conflated information about another person mentioned in an article about the restaurant bankruptcy with the actual owner. If I didn’t look at the source I would have believed a completely untrue statement, which had not been definitely stated anywhere
Excellent, thank you!
Thank you for this.
Quite apart from the water, land and energy use and abusive employment of trainers inherent in AI, it´s also based on the unpaid use of everything every written or photographed. AI as most people use it is a way to make a few people rich based on: outputs not grounded in truth or a first-person experience of the world.
We´re here to talk to other humans not Claude, ChatGPT etc.
Here here! I fully support your decision, and I am glad you made it.
GIGO is an acronym dating back to the earliest days of computing. It stands for “garbage in, garbage out”. It’s still true, but now it is happening at an alarmingly increasing rate.
“A lie gets halfway around the world while the truth is still putting its boots on.”
“AI gets halfway around the world while the truth is still putting its boots on.”
There, fixed it for you!
Excellent! This is the attitude that attracted me to CC in the first place. Even before the advent of AI so many other automotive websites seemed childish in tone, with a lot of pointless partisan squabbling and fan-boy nonsense presented as fact. While I have used AI to assist research for writing fiction, I steer clear of it for serious matters. It’s simply not trustworthy as a research tool.
In my line of work, I deal with AI as a matter of daily tasks. I’m a heavy user, and also need to be a reviewer, both for professional and personal reasons. In my experience, there are a few things to be thought about (I won’t say “as a user of such and such I hereby declare”).
1) As has always been true of information systems, GIGO (Garbage In, Garbage Out)
2) Most tools (In information systems and other technologies, beginning with a hammer) are tremendously powerful as long as they are used by people capable to do so and FOR the means they are intended. A hammer can be deadly.
3) I can’t speak about the usability or capability of tools I’m not prepared to deal with. If I were given a machine (say a gearbox, or a front end, or…) and a set of screwdrivers, declampers, clippers, and whatever other things are needed to dismantle, fix and reconstruct it, the device will remain as is in the best case, in the worst case I’ll probably ruin the chance that a capable human will repair it.
4) Last but not least. CC is known and followed by many people, and many of those, counting myself, are here and subscribe because of its outlook, serious people writing from experience, sources and knowledge, but also because you always feel there is a human behind. I’d say, you’ll never be able to prepare as many articles from as many people telling you a human story that happens to be much related to an automobile, if there weren’t enough human beings, real ones, devoted to do the task.
So, Paul, your explicit decision is very welcome. I, for one, like to leave work at the office, and though I play with AI to find out whether a Porsche 356 could be ordered with a column shift lever, I’d rather trust my website of choice. Thanks a lot for all your effort, and that goes for all the team also.
Should we said goodbye to “phantom” cars and others AI generated cars that never were like this 1975 Packard Patrician and 1975 Studebaker Lark in the “what if?” and “what might have been” categories?
https://www.curbsideclassic.com/blog/what-if/what-if-1975-packard-patrician/
https://www.curbsideclassic.com/blog/what-if/what-if-1975-studebaker-lark/
I am very glad CC only uses 100% organic intelligence.
I use AI for analysis and summaries but I always check sources first and CC is often a primary source
Well, I guess you answered my question about AI use in C C articles. can understand the ‘accuracy level’ of AI giving false facts about specs like engine size, model nomenclature, release date, prodution numbers and other facts about the cars discused here. My question was if I can use AI generated ‘cartoon’ strips in multi panel formats to provide a weekly cartoon for readers of CC. They will be of humorous, semi satirical, seasonal and nostalgic nature.And since all my egenerated images are done many times until I get them right, they are not slop.I can provide images for your approval if you give me a chance. Thank you.
As I mentioned in the post, yes, we are open to posts that use AI-generated images for precisely the reasons you stated. We’ve done that in the past and had a lot of fun with it. AI can be a great tool for bending reality visually in a humorous way. Feel free to contact us directly at curbsideclassic(at)gmail.com
Thank you Paul. This is a good thing. AI should not be trying to write ‘fact based’ articles.
As an example, I was on MSN the other day, sadly a default website on a browser I almost never use and got sucked in by a click-bait article (one of many tailored to my interest in all things automotive and all things aviation). The article was about Mustangs, and it was a link to what I thought was a halfway respectable site called Slash Gear.
The article mentioned the S-197 Mustang (something about which I know, considering my handle here), and that the base engine in 2007 was a 4.0L V8… um, I was just under the hood of the car this very morning and only counted 6 spark plug wires. Hmmmm….
Me thinks the author used AI to gather his so-called ‘facts’ and blew this one big time.
This isn’t the first time I’ve seen this. AI gets aviation details wrong all the time too.
I can confirm that a rapidly growing number of car websites are using AI to one degree or another. They hire young kids for peanuts (often in faraway countries) that have no real historical context or experience and expect them to generate content that will pop on the web. AI slop. Reader beware!
All you need to know is the disclaimer starting to show up: “AI-generated content can contain errors.”
Agreed. AI is hardly “super intelligence”. At this point, I am thinking more that “AI”, at least when it comes to historical information, is the worst sort of “I believe it because I found it on the net”, sort of intelligence.
For reasons not germane here, I have been researching the military situation in the Philippines, just before the war. When I asked the AI agent why rifles the Philippines bought from the US were not cut down to carbines, to be more manageable for the average sized Filipino soldier, the agent said they could not be modified because of the Neutrality Act. I told the bot the Neutrality Act did not apply, because the Philippines were a US possession. So the bot pivoted to a different argument. It insisted that there was no armory in the Philippines that could have cut down the rifles. Wrong again. The Philippine Ordinance Depot certainly had the capability, because, some years before, that Depot had cut down rifles of a different model, so they would be more manageable, for the members of the Philippine Constabulary. The bot has told me other things I know to be haggis, just in this one project.
I need to find an in depth book about the P-40, because I don’t trust what the bot was saying about the armor.
Tienes toda la razón del universo. La IA está afectando a las publicaciones no sólamente de automóviles, también de todos los temas y rubros. Es muy triste que cada vez que yo pongo el buscador determinado tema, se me aparece una respuesta de inteligencia artificial sin que yo se lo haya pedido , ésto implica que están plagiando todos los grandes trabajos que los autores científicos y amateurs han hecho con tanto esfuerzo. Ésto está poniendo en peligro la diversidad y los matices de abordaje que son tan interesantes de descubrir. Yo te escribo desde Argentina y amo Curbside Classic desde hace unos 10 años . Tengo mis puntos de vista y me encanta compartirlos con gente de la cultura estadounidense, asi como jamás podria burlarme de los antiguos automóviles soviéticos porque detrás de cada automóvil hay una arqueologia industrial , culturas, sacrificios y sobre todo mucha carga emocional de cada país que se dedicó a ensamblar sus vehiculos con personalidades adecuadas. Sé que Curbside Classic es la ÚNICA publicación digital que siempre fue pluralista, genuina, que siempre recibe con entusiasmo los testimonios de países lejanos como el mio. Te saludo con mucho afecto desde Argentina, no existe inteligencia artificial alguna que pueda confundir a la gente que prefiere el buen discernimiento. También le mando un abrazo a todos los lectores de Estados Unidos , porque son la fuente genuina de todas estas discusiones, porque entre todos uncluidos los extranjeros queremos que Curbside Classic conserve su cordón umbilical con el público, que tanto te agradecemos tus fascinantes articulos que abarcan todas las culturas del sector automotriz
Thank you Paul.
I will go against most of my “R-word” compadres and say I think Ai shoulda been stopped at the outset and preemptively put out of our misery. No way it ends well.
As for the LHS, I recall reading in a car magazine back then that it was Luxury High Series.
On a related note – I did own a 24V loaded version of the Concorde, which I pimped out with 300M alloys. It was quite a fine car. Later kept it going with a newer 300C Hemi (with straight pipes!) that was even better.