The "Lost in Translation" Location Trick: Letting AI Find Your Contribution Location For You

[abermans]

You may know the scenario: brilliant photo of a shopfront or sign abroad, and you have no real idea what the place is called or how to find it in Maps once you’re back to upload it.

Normally you’d lean on your camera’s geotagging to help Maps suggest nearby places to attach the photo to. Useful, but not a full fix.
In a country with a different lanuguage, that suggestion list may just a column of names you can’t read and can’t match to the shopfront in front of you.
And just as often, the actual spot you want to credit isn’t on that list at all, so you’re stuck searching manually, which means you need to already know its name. Which is the entire problem.

Turns out there’s a stupidly simple fix sitting on your phone already: Just ask an AI chatbot with image and maps capabilities (I use Gemini, I assume others may do it as well) to identify it for you first.

Why this works

These tools don’t just read the text in the image, they seem to recognise the storefront, the signage style, the architecture…and cross-reference all of that against what they know.
In my experience testing this repeatedly since getting back from Japan, the accuracy has been genuinely impressive. It’s not “somewhere in Tokyo” vague, it’s “here’s the exact street, here’s the pin, here’s the name of the business.”

The actual steps

Nothing complicated here, which is rather the point:

1. Open the photo in your gallery.
2. Hit share, and send it straight to your AI chat app (Gemini in my case).
3. Type a simple prompt: “Find this place in [city/area] on Maps.” Adding the rough area helps narrow things down if the location is common (a chain, a generic-looking street).
4. Wait a few seconds for the response.
5. The AI will usually name the place and can often pull up a map with a pin directly in the chat.
   Gemini actually allows to open the map location and take it from there.
6. You can copy the name and use it in maps in cases location does not open.

That’s it. Photo in, location out.
Yes, it will should also work quite well on location for which you know the name and language but please don’t be that lazy

[MortenCopenhagen]

I use Google Lens for this. Works pretty good too.

[AdamGT]

Great tip @abermans and an interesting applicatoin of AI. Would have been handy quite a few times when I was uploading a backlog of photos from my travels. Do you think this method has a greater chance of success in more populous locations?

[WilfriedB]

@AdamGT

Do you think this method has a greater chance of success in more populous locations?

Very likely, @AdamGT and @abermans. Don’t forget, AI can only repeat what already can be found on the Internet. So, I assume it compares to existing photos or am I wrong?

I did use Gemini successfully in a similar way:
We often take photos of sculptures or paintings in museums or other in the public. If I don’t forget, I also make picture of the description to be able to tag my photos with title and artist’s name. Sometimes, there is no title or I forgot to make the photo. Using Gemini on the desktop, you can drag the photo and drag it into the browser session. Gemini not only tells you title and artist, but also some more background information, you wouldn’t easily find where you took the photo.

[abermans]

Yes, it seem to be working quite well.
I tried to “reduce” the help and checking a place that have multiple branches in London and it correctly identified the location (maybe it is smart enough to assess location tag? Who knows - these tools are becoming smarter by the hour).

I assume the party of free AI tokens (usage) may not last forever but for now and my need it’s good (I got 100s of such photos from Japan which I’ll use this trick for).

At some point I can see this contribution effort fully AI automated…

[AdamGT]

It’s a very interesting and useful application of AI @abermans I also like the application that @WilfriedB describes and can see similarly examples of where this would help. Great stuff!