There are more conflicting statements here than necessary. This is not the case: I do not need to examine the source code in order to assess the privacy of my data. There are multiple ways to ensure that data remains private and secure when cloud connectivity is not mandatory. One of the simplest, most effective, and most fundamental approaches is to take the system offline. While this method is also the least convenient, it is often the most reliable.
You state that “the privacy of a connected cloud system is really no different from proprietary code (the only difference being that you can unplug one).”
It is important not only to mention but to emphasize that there is no such thing as absolute privacy in cloud-connected systems. At best, users are given an illusion of safety. Comparing a cloud-dependent system to a closed-source offline application is therefore misleading. The difference you acknowledge is not minor—it is fundamental. Saying there is “no difference” while simultaneously pointing out a difference is logically inconsistent.
To give a simple example for non-IT audiences:
If a system can be taken offline and still function as intended, its privacy is limited only by its physical security. In this scenario, whether the software is open-source or closed-source is largely irrelevant. Similarly, if that same system is placed behind a properly secured network with stateful inspection and firewalls, its security is defined by the strength of that network.
In contrast, if a system is cloud-dependent, the user cannot fully determine its security or confidentiality. In many cases, reviewing the EULA makes it clear that cloud providers are required to share data under certain circumstances. There are no exceptions to this reality.