Oh, hey, a question I can answer! I'm not a historian, but I do work with airborne remote sensing and GIS professionally.
Yes and no. All maps are imperfect and have some errors. Partly this is because there's uncertainty in measurements, and partly it's because the Earth's surface changes over time. Also, any map will be incomplete because it will have finite resolution. Details are omitted out of necessity. So in that sense, all maps contain errors that need to be fixed, and airborne mapping data is one source of information for addressing those errors. A modern aerial LIDAR survey can provide vastly more topographic detail than if you were following the Manual of Topographic Methods, for instance.
However, assigning a precise location on the Earth's surface to any part of an airborne image is a difficult problem. You need to know where the plane is, what direction it's pointing, and how close everything it's looking at is to the camera. These days that's mostly done with GPS, inertial measurement units, and digital elevation models. But prior to the advent of GPS, it mainly meant identifying features and landmarks in the image whose location you already know from ground surveys and using those to tie the image to existing maps.
It's also worth noting that whenever we're discussing "location" and "elevation," it's all relative to something. Pre-GPS, this meant surveys were relative to a fixed point, such as the Meades Ranch Triangulation Station that formed the basis of the North American Datum of 1927. Specifically, everything is measured relative to that little dot in the center of the disk. So ground surveys were actually quite capable of a high degree of precision and accuracy. Even now, if I really want to know where something is, I'll want a ground survey. With GPS, that means setting up a GPS antenna over a benchmark for several hours, or ideally several days. Often, such an antenna is set up over a known site while your airplane is flying. Then you compare the position of the plane relative to the antenna rather than simply relative to the GPS satellites within its line of sight. This can be used to correct some positional uncertainty. But even then, the compounded uncertainties mean that the airborne data can't reach the precision and accuracy of a ground survey. So a modern airborne LIDAR survey will provide a great deal more detail, but the location information of select points may be worse than what can be obtained on the ground, even using older methods.
What airborne (and satellite) mapping really gets you, and why it came into widespread use, is speed, breadth, and access. If you're a British commander in WW1 and you want to know what the German trench network looks like, well, Jerry's not about to let you waltz over there with a Jacob staff and a Theodolite and take notes. But you can fly overhead and get a pretty good map from aerial photography, and you can repeat that flight a few weeks later to see what's changed.
And while it took 45 years for the USGS to complete its first set of 7.5 minute quadrangle topographic maps, the Shuttle Radar Topography Mission took 11 days (plus post-processing), and covered a much larger area. And getting the level of detail from a ground survey that you can get with a LIDAR survey is simply impractical and would be prohibitively expensive.
I hope that answers your question!
