I understand that you are prepared to do the work necessary for searching the literature, but that you are looking for tips that avoid swimming aimlessly and wasting time going down unproductive lines. In principle, this is resolved by having a clear idea of what you are looking for. That way, you don't waste time exploring interesting ideas that are not immediately relevant. (And since it sounds like you are intellectually curious, then almost everything is interesting.)
I have found that one of the most valuable things to focus on in a literature review is theory, which I define as an explanation of the relationship between two or more measurable concepts. So, what you can focus on in a literature review is:
- Clarify the theoretical concepts that interest you. A theoretical concept is necessarily something that is measurable, whether qualitatively (e.g. Type A versus Type B versus Type C) or quantitatively (i.e. counts, dates, etc.). This is the absolutely most crucial step to identifying theory.
- When you search the literature, you should look squarely for these theoretical concepts. You might not know the measures that exist in the literature at first, but at least you know that you are looking for the parts of articles that talk about measures of your concept of interest.
- You ignore anything that does not talk about measures of your concepts of interest. This means ignoring articles that don't talk about your concepts of interest at all, and it also means that even if an article does talk about your concept of interest, you skip the parts of those articles that don't talk about the measures of these concepts. (That said, of course, you should read anything that helps you understand your focal concepts better, even if they don't talk about measures.)
For example, if you want to study conditions that make people feel threatened by robots versus when they warmly accept robots, the key concepts would be:
- Concept A: Type of robot interaction
- Concept B: Human reaction to the robot
With these concepts identified, then you would search for literature that describes:
- Different types of robotic interaction (e.g. personal voice assistant versus hardware robotic device; "machine-looking" robot versus humanoid robot; robot assists work tasks versus robot replaces human job; etc.)
- Different ways of consideirng or measuring human reactions to robots (e.g. a scale ranging from violent rejection, such as vandalism, to warm acceptance, such as using the robot many times every day)
The results of the literature review would be a theory that explains how and why certain types of interaction with robots provoke various human reactions. (The theory would necessarily need to be enriched with context such as human personal characteristics, legal frameworks, etc.)
What I gave is a very brief summary and illustration of a literature review methodology that I am still in the process of developing, but I have an older working paper that describes the core principles in quite some detail. Even though I have unpublished ideas on this that are much better refined, the section in my older working paper on searching the literature explains this approach that I summarized in more detail.