February 2nd, 2026 - Down a ChatGPT Rabbit Hole for Project PRO.2 and its History

February 2nd, 2026 - Down a ChatGPT Rabbit Hole for Project PRO.2 and its History

As I noted in my blog post here: Some Internet Sleuthing to Uncover Project PRO.2's past in Japan, a lot of information about my car's history has been uncovered directly from changing technology over the past 21 years I've owned her. 

• Virtual Machine/Electronic Parts Catalog: I used a Honda Japan EPC to determine some of the factory options that were on my car when it was built, such as Honda SuperSound stereo system, etc.
• Plain Old Telephone System/Email: I obtained some information via plain old telephone and email when I called Honda Japan.
• Google Translate iOS App: I was able to glean information from the Google Translate app on my phone being able to translate the Kanji using my iPhone. 
• This allowed me to translate the service stickers on my strut tower and airbox to get even more history
• Allowed me to translate the pricing matrices from two Japanese Honda Verno stores to more accurately re-create the original sticker price
• CarVX Report: This allowed me to obtain information about my car that pointed to it never having been sold to the exporter via an auction house.
• Japan Car History Check Website: They were able pull information from Japan's Ministry of Land, Infrastructure and Transportation which showed me how many previous owners there were and where the car was registered while in Japan. 

And this past weekend I went down a rabbit hole with ChatGPT about my car's history. Using Microsoft's Copilot extensively for work, and ChatGPT for personal use, I have a healthy distrust for AI in general. I know that AI - ChatGPT especially - can hallucinate and present findings and analysis that is geared more to what you want to hear versus what's factually correct. Believe me when I say, I grilled it for hours, asked it for its sources, asked it to confirm findings, challenged it when it went down a path I didn't think was warranted based on the evidence presented. I asked it to ignore the previous inquiries I'd made and start absolutely fresh to confirm I hadn't accidently guided it one way or another.

I did everything I could to disprove what it was telling me. I didn't want to come here and misrepresent something that was incorrect because I trusted AI. 

I started with some conventional Google searches but kept running into roadblock after roadblock, mostly due to the language barrier. When suddenly it occurred to me. Why not ask ChatGT about the dealer location, and also this whole first owner being listed as the dealership. It's always struck me as a little odd, so maybe AI can help me.

After a lot of back and forth, I uploaded the MLIT registration report I got from Japan Car History Check and asked it to do an analysis, ignoring all previous queries and analysis I'd already asked it because I felt like there was an effort being made to tie the responses back to previous queries. This was the reply. I felt comfortable that we would be dealing impartially. 

Understood. I’ll do this as a true reset in reasoning, not just rhetorically but methodologically. I’ll ignore all earlier conclusions, locations, dealer names, guesses, and even Honda-specific narratives unless they are strictly unavoidable facts.

Think of this as: What can an outside researcher conclude if they walked in cold and were very conservative?

---

Ground rules for this analysis (clean-room)

• No reliance on earlier addresses, maps, or dealership assumptions
• No back-projection from modern Honda Cars entities
• No reliance on anecdote, enthusiast lore, or “this is usually how Honda did it”
• Only conclusions that would survive scrutiny from:
• an archivist
• a historian
• or a corporate legal department

After over a week of back and forth messages to ChatGPT and countless hours spent doing my own research and exhausting every possible angle I could think of that we could explore, I was finally satisfied that I was getting the most accurate answer I was going to get from ChatGPT and that in order to take the investigation further, I'd need to reach out to other sources for additional information/confirmation. 

So it appears that it's possible, and highly likely, that my CR-X began its life as a dealer demo, test drive unit or possibly even service loaner. It wasn't sold to its first retail customer until November of 1991. This is a bigger plot twist than when I discovered it was never sold to the exporter at auction as I'd been told when I bought it in 2005. 

Note: I continue to say "possible" and "possibility" because I've not been able to confirm these things beyond a reasonable doubt, and I'm still not ready to fully trust our robot overlords just yet. As a matter of fact, after posting about my adventures with ChatGPT, a friend from Japan has already mentioned that it's unlikely that a rural dealership such as Hachinohe would have used a CR-X as a service loaner. 

Here's what we discovered:

Dealership Location

• I started by asking ChatGPT (CGPT) to help me find the Honda dealership that originally sold my CR-X: Honda Verno Aomori, the dealership Honda Japan told me was closed. 
• I asked about specifically about a Honda dealership in Hachinohe City which corresponds to my Ministry of Land, Infrastructure and Transportation (MLIT) registration search from Japan Car History Check (JCHC).
• CGPT says that many locations kept the same land and buildings. They changed legal entities, signage and official records.
• When Honda Japan says “closed”, it often means “This legal dealership entity no longer exists under that name.”

• Hachinohe City as the retail location makes sense because Hachinohe City was (and still is) a major population and industrial center in eastern Aomori and one of the primary market areas Honda targeted when siting Verno dealers.
• A sanity-check on Wikipedia confirmed that Hachinohe's population as of 1990 was approximately 247,000 people. Definitely not a major center. 
• Honda strategically placed Verno stores based on market size, not just prefectural capitals .
• So “Honda Verno Aomori (Hachinohe)” is entirely plausible.

Gruppe Professional Golf store 青森県八戸市城下3丁目15-25

• I checked the Google Street View armed with some architectural clues CGPT provided to try and determine whether the dealership could have been an old Verno dealership prior to the Honda Cars rebrand of the early 00s.
• Look for: Older, low-slung dealership buildings, lots dating visually to the 1980s–1990s.
• Even if the building is modernized, the land parcel is often unchanged.
• The location I found on Street View was currently a store called Gruppe Professional Golf shop. 
• I was skeptical that the location used to be a car dealership from the look of the building at a glance. 
• At first glance, I thought maybe it could have been a head office for a dealership group or something.

• Looking at the street view from a different angle, I noticed something - garage bay doors. 
• I asked CGPT to analyze a screenshot I took Google Street View and it came back with some clues I'd missed initially:
• The large, full-height glass frontage that has a wraparound so that cars could have been seen from multiple angles.
• The service bay wing with multiple wide roll-up doors
• Red horizontal band wrapping the building is in exactly the spot where Honda Verno signage would be incorporated.
  
  
  
  
• Next I looked at the Google Maps overhead view and it became clear to me that this was a lead worth following up on. 
• The collection of residential houses sits in exactly the spot where a large, open parking area for new vehicle inventory would be kept.
• Subdivision of the property after its purpose as a car dealership came to an end was likely.
  

• Earliest Google Street View archive was from 2011 shows it was the golf store at that time, so that was no help. 
• I checked the Company Profile on Gruppe's website shows that they purchased the store in 2009.
• I knew it was a long shot but I called the Gruppe store, and using iOS 26's translation matrix, I was able to have my telephone call translated in real-time. I asked if they happened to know the purpose of the building prior to 2009 and whether it was a Honda Verno store. 
• The person on the phone couldn't help me. It was a long shot. 
• I checked the GIS aerial survey photography from that time period it would have been a Honda Verno dealership. Here's where it got interesting:
• I found archival aerial photography from 1981, 1984, 1987, 1992 and 1997.
• The resolution of the online photos isn't great, and ironically the best photograph was taken in 1981.

• You can make out the showroom and the service bays in the 1981 picture, and clearly the dealership parking lot for outside inventory and parking. 
• Turns out that the houses were there all along and that the dealership lot extended in a different direction than I'd expected. 
• Today, the large structure near the point on the left is still there, but the parking lot is divided by two fences. A sign on the fence nearest the main road states that it is a parking lot for rent.
• The telephone number on the signage doesn't appear to be registered to a business.

I even had ChatGPT render the screenshot from Google Street View to show what the store would have looked like in 1989 as a Honda Verno store.

Confirmation From Independent Sources

• Even though I think it's safe to say I've confirmed that this was, in fact, the same dealership I've been searching for, I still have some questions.
• Some of the next steps to explore is to try and locate the following information from the Hachinohe library or the National Diet Library (国立国会図書館)
• Nippon Telegraph and Telephone (NTT) telephone books of the area from 1990, looking for Honda Verno stores, including any mention of  head office or service center etc at the associated address from my MLIT documents.
• If they have available any digitized newspapers from September 28, 1989 with possible car dealership sales advertisements that could assist.
• Edit 1: I received a reply from the Hachinohe library:
• They stated, in Japanese, "We are sorry, but we do not have any old telephone directories in our collection, so we are unable to confirm your inquiry."
• "We do have residential maps in our collection, and we have found that a 1988 map shows a Honda Verno Aomori Hachinohe Store at 15-25 Shiroshita 3-chome, Hachinohe City, and a store with the same name in the same location can also be found on a 1990 map. (We do not have a residential map from 1989.)"
• I have now confirmed, independently, that the store currently owned by Gruppe Golf was once Honda Verno Aomori (Hachinohe Store) according to a 1988, and a 1990 map, stored at the Hachinohe Library. 
• And since the MLIT report from JCHC shows that address for the User of my CR-X, I can now say with certainty, that is the address where my CR-X was first sold. Now I just need to confirm... to whom?
• Edit 2: Further correspondence with the Hachinohe library has taken place. They have offered to make physical copies of the pages from the 1988 and 1990 maps and I've confirmed that they can copy a limited number of pages from the local newspaper. 
• I have chosen to request reprints of the front pages of the morning and evening edition of the 東奥日報 (To-o Nippo) newspaper on that day as well as any specific mention of automobile advertisements for Honda Verno Aomori or the Honda CR-X. If they found none, I would accept printed pages for any advertisement from Honda Verno from the time. 
• I'll have to pay a modest copying fee and payment will need to be made via Bank Transfer. I'll come back and update when that's completed.

Registered Owners

• Checked the MLIT registration document I received from JCHC. 
• The address of the first registered owner, listed as 株式会社ホンダベルノ青森 八戸店 (Honda Verno Aomori Hachinohe Store) was 青森県八戸市城下3丁目15-25  (3-chōme-15-25 Shiroshita)
• This address is therefore not a residential address, obviously. 

• Uploaded the JCHC MLIT registration history into ChatGPT and asked it for an analysis.
• According to ChatGPT, Japanese vehicle registrations distinguish between two key roles:
• Owner (所有者): The legal title holder
• User (使用者): The party who actually operates and stores the vehicle
• This distinction exists specifically to support leasing, financing, and corporate ownership structures. 
• The User is often the best indicator of how a car was actually used day-to-day. 
• Japanese automakers often registered new cars under a leasing subsidiary rather than directly to a private buyer.
• Legal owner (often finance company)
• Actual user (customer)
• This structure is repeatedly visible in:
• Late-80s Hondas
• Early-90s Toyotas and Nissans as well.
• A ~4-year owner-retained period is entirely normal and appears in multiple verified cases.

• A careful look into the first two registrations of my CR-X indicates that:
• September 28, 1989
• The first registered owner was Honda Lease Co. Ltd
• The user was Honda Verno Aomori
  
  
• November 8, 1991
• The second registered owner was Honda International Sales Co. (HISCO)
• The user was a individual person.
  
  
• We've already extensively searched the first registered User's address in the searches above. It belonged to Honda Verno Aomori - Hachinohe store.
• A Google search of the second registered User's address from the MLIT report indicates a residential address.
• CGPT could not find any evidence that the second registered User's full name was affiliated in any way with Honda Verno Aomori as a corporate executive or employee. 
• That being said, CGPT did admit that its ability to search such information was limited since the dealerships were privately held businesses and in Japan it's not common to publish the names of dealership employees.
  
  
• A ChatGPT check on HISCO and what could be interpreted from corporate records in Japan revealed that:
• Was formed in the early 1970s as a sales and distribution company connected with Honda’s broader sales and trading operations. 
• This suggests that the name Honda International Sales Co. was used historically by Honda for certain sales or distribution activities, potentially including handling vehicles on behalf of the company in contexts such as export, wholesale, or resale.
• A ChatGPT check of Honda Leasing Co. Ltd revealed that in the 1990s
• Was a back-office financial services entity (not retail sales)
• Honda Lease Co. Ltd was used to hold title to vehicles while they were used in dealer demonstrations, promotions, or internal fleet use.
• This legal ownership by Honda Lease does not imply retail sale to the dealer; it reflects Honda’s internal financial structure of the time.
  
  
• I advised CGPT that the model information for the Civic and CR-X, specifically the generation change between the 1991 and 1992 model years.
• Cross-referenced those dates with the timeline for the sale of my CR-X from Honda Verno Aomori to the retail individual line up almost perfectly:
• The 92-95 Civic launched in Japan in September, 1991
• The CR-X del Sol launched in Japan in March 1992, after production of the 2nd generation CR-X ceased in February. 
  
  
• ChatGPT drew the conclusion that my CR-X was most likely
• Used as a dealership demo, service loaner, or possibly some sort of promotional vehicle.
• A friend from Japan on Facebook saw a post I'd made and let me know that in rural Japan, back in those days, it was highly unlikely that a dealer would have used a CR-X as a service loaner.
• Was sold to a retail user via HISCO in November 1991 at the end of the EF8 CR-X's marketability as the new generation Civic was already being sold and the new CR-X del Sol was on the horizon in a few months' time.
• Evidence used to make this conclusion:
• First user was listed Honda Verno Aomori
• First owner was Honda Leasing Co, Ltd
• According to Japanese tax law from the time, companies were able to write off the depreciation of leased assets as well as the lease payments similar to North American tax law.
• The sale to the second Owner/User aligns with the end of the 2nd generation CR-X's relevance in the sales/marketing strategy.
• The second owner being transferred to HISCO aligns with the transition from corporate lease to retail ownership financing.
• The second user's address being a residence as opposed to a commercial address like the first user.
• Anecdotal information from me that I was unsure when the EF8 ceased manufacture but that I was aware of EF6 models being produced well into 1992. 

Attempted Confirmation From Honda Japan

• The next step I took was to email Honda Japan for confirmation.
• The email bounced. I called Honda Japan to get the new email address and I updated my earlier blog post accordingly.
• I had asked Honda Japan for confirmation on the following:
• Whether the vehicle was originally designated as a Honda Verno dealer demonstration, showroom, or promotional vehicle during its initial period of use under Honda Leasing and Honda Verno Aomori.
• Whether the transfer in November 1991 to Honda International Sales Co., Ltd., with a private individual listed as the user, would be consistent with a retail sale financed or administered through a Honda-affiliated company at the end of the EF-series CR-X lifecycle.
• The address and exact location of the Honda Verno Aomori dealership in Hachinohe that handled this vehicle in 1989, as it existed at that time.
• Two days later they replied and told me that they weren't able to confirm or deny any of the following information. It had simply been too long and records were not available from them. 

Attempted Confirmation From Japan Car History Check (Anecdotal)

• I don't explicitly trust ChatGPT's analysis of what the norms are for Japanese vehicle registration. In general, I don't fully trust AI, period, but I do know someone who I would consider to be an expert in reading Japanese registration information.
• JCHC would have is a specific knowledge of Japanese car registrations and could possibly help me with some pattern recognition when it comes to how the first owner/user usually appear.
• They wrote back saying while every situation is different, but it is not uncommon to have a Honda dealership show up as the first registered owner, the user is typically the retail customer. 

Other Anecdotal Testimony

• A friend on Facebook who lives in Japan and drives a CR-X also confirmed, anecdotally, that the way my car was registered is not normal, and it was likely due to it being a demo car. 
• He also admitted that it seemed too early for it to have been sold to a customer (September 28th, 1989).
• He also mentioned that if the dealership held the car for more than 6 months or so before transferring the car to a retail customer, that would further strengthen that position it was a demo. If it were a retail sale, the retail user's name should appear first in the MLIT report.
• He told me about another CR-X owner we both know who had a registration history check done. 
• That person's CR-X had an ownership history that began with a retail user, but the second owner/user was a dealership. 
• Likely a retail user trading it in and the dealer keeping it for their use for a time. 

────────────────────────────────────────

As more things are uncovered I will come back and update this. 

 

Remember, this blog was written in chronological order so, you will need to click OLDER POSTS to move forward in time, and click NEWER POSTS to go backwards.