That is exactly the problem. Windows does not do record locking. It only does FILE locking. Any finer granularity is up to the application. For a passive Access back-end file, Windows connection data stops at the file's "boundary" - i.e. the raw inbound connection, from Windows's point of view, is to the WHOLE FILE. Only the device drivers of the machine hosting that DB know exactly where within the file there is any activity, and that knowledge is ephemeral at best. Access itself tracks that information, which it ONLY does for very short intervals - particularly short if you have Optimistic Locking rather than Pessimistic Locking. (And it doesn't track it at all if you have No Locks.)
The question being asked may be impossible to answer for a "pure" Access back-end file because all of the "real" connection data is in the Access workspace, a memory structure inside an Access process, except for whatever record is kept in the .LACCDB or .LDB file. Assuming you could even identify the computer having another connection, and THEN assuming you could track that back to another process, you still have a security issue.
The problem with asking another process where it is connected is that Windows security rules will not let you touch the inner workings of an external process. For those who remember the old military "Orange Book" that sets security standards for operating systems that are approved for U.S. government use, the C2 standard includes "inter-task isolation" as one of the four critical requirements. A newer and more stringent standard exists now, but it still includes that requirement.
Therefore, the ONLY way for anyone to EVER find this out for Access is to write the app to leave traces when a connection is made because otherwise, you cannot get there from here. Now, if we are talking about some active SQL engine on a server, the active server system can track that for you better - but again, only if it is set up for that. The problem in a nutshell is that, in general, for most operating systems I know, you CANNOT find out what someone else's process is doing unless it is set up in the first place to tell you what it is doing. Even with privileged processes running as SYSTEM, there is the problem of knowing WHAT to ask and HOW to ask it - and catching it at the right time for the info to be available.
This idea of having the app leave traces is exactly in line with my "Old Programmer's Rule #2" - which is that Access won't tell you anything you didn't tell it first, or at least tell it HOW to tell you. If you are going to require usage information then remember to tell Access how to TELL you that information ahead of time.