When I first got onto iNaturalist, I imported pretty much everything I got on Flickr from the last 10+ years or so. Uploading old pictures is fine and can add valuable data for e.g. phenology studies going back in time, as long as you still know the date and approximate location to put a pin on the map.
I am working on putting in old observations. We have a wealth of current data, but not a lot from earlier times and even less from before digital cameras became commonplace.
Remember if you don’t know the exact place, the accuracy radius can be expanded to cover, say, a whole state park - there are times when I noted where I was, but not very accurately.
I’m trying to decide when an accuracy radius is too large to be helpful. I’ve seen some covering half a continent and I don’t think I would include such an observation. Of course, it may have been an accident, so I try to leave a message for that person.
What kind of radius do others set as their limit for usefulness?
I add my hiking photos each week. Am slowly working back, with the help of my Life List, and adding extra species from earlier hikes. Also working my way around documenting what lives and grows in my garden. And then even earlier photos, from my blog and travelling.
PS for other identifiers. I reset to Observed today, not uploaded today as that gets confusing for seasonal observations.
There’s also a - What is your oldest photo - thread. Including some inherited from earlier generations.
If you experienced it, you can post it - go for it!
I would just echo some of the comments of other posters to be sure to accurately reflect any uncertainty in date/time and location in the observations.
If you aren’t sure of date, you can still upload without one (though it will be casual), but this can be useful for adding something to a personal list.
For location, be sure to set the location circle (accuracy) to be wide enough to encompass all of the potential locations that an observation might have been made in. People using the data would generally rather filter out data with large uncertainty to have a slightly smaller, higher quality dataset than have a dataset that has incorrect location data.
I am now getting physical slides scanned. I’ve posted a few. (It will be a long job.) They go back to the 1980’s, and I’ve even got a few from my grandmother from before that. Making sure the time and date are accurate takes some effort, and large accuracy circles are often necessary. However, I think they’re worth it – sometimes they record things that no longer exist. So, in answer to your question – go for it, and you’ll have company.