Since I have been elected a moderator at Artificial Intelligence SE (about 1 month ago), I've noticed that certain questions that have high value (for our site, but not only), such as What are the AI technologies currently used to fight the coronavirus pandemic?, have not become hot questions, while others, which aren't as socially relevant (or relevant to the goal of our community) as the former, such as this one, have become hot questions. In the past, I've seen this pattern multiple times, not only on AI SE. For example, I've seen this on Stack Overflow, even recently, although I can't remember the exact title of the post, the post became hot because there were two successive statements (of the form k++; and k--;) that canceled each other. Of course, this was just a mistake of the writer of the code (i.e. those statements don't have any effect!!). Is this really a question that deserves to become hot?
This clearly suggests that the current algorithm or heuristic to determine which questions should become hot (or not) isn't working appropriately. Although I am aware of the ability that moderators have to remove a question from the hot questions list, I think this is not sufficient, in the sense that, in the case certain questions are not even considered for the hot list, they should be manually added. Furthermore, note that changing the hardcoded heuristic/algorithm doesn't improve much the situation, because, often, e.g. in the case of this coronavirus situation, the heuristic needs to change flexibly and adaptively, according to the situation, and, currently, only humans can do this (intelligently).
Therefore, I suggest the following alternative approach. There should be a separate voting system (possibly different than the current voting system, which mainly grades the quality of the question only according to the users that have seen and read the question or are specifically interested in that topic) that can be used to explicitly vote or promote a question to become a hot question. More precisely, I suggest that mainly moderators (and maybe other users) could vote for a question to become hot.
This can be especially useful for questions that don't receive answers, although they are clearly important. Furthermore, by doing this manually for several years, we could collect "good data" that could be used in the future to train a machine learning model to help to recognize actually potentially hot questions. Finally, I think that the community should decide which questions should become hot. A healthy community should be able to decide which questions actually deserve more attention.
You could argue that this new voting system is exactly similar to the existing voting system. Well, I don't see a big problem here anyway. Votes really represent the importance of the question, but, in this case, they will specifically represent something else (e.g. "Should this question attract more attention?"). In any case, I believe this will help to promote certain questions that should really be promoted to the hot list that are not currently being promoted.
This voting system would be more similar (in scope) to the voting system used to close or re-open a question rather than to upvote/downvote it.
Some people claim that this system may not scale for sites with many questions (e.g. Stack Overflow), but this is a poor argument because people already use voting systems to close questions for other types of reasons (e.g. off-topic, too broad, etc.). Furthermore, I am not saying that this voting system needs to necessarily replace the existing system. Maybe it can simply augment it. This would be an option that moderators (or other users) would further have to deal with certain relevant questions that the underlying heuristic didn't even consider (for whatever reason, but probably because it wasn't intelligent and adaptive enough to understand it).
