I propose the ability to request ad hoc review queues - to highlight a problem that requires a concerted effort to solve, but doesn't occur often enough to merit a permanent review queue. It would require a clear description of the problem, a list of proposed fixes, and a certain threshold of posts to be considered relevant (say, 1000 questions/answers). Once emptied, the queue could either disappear completely, or remain dormant until that threshold was reached again.
This could happen simply through a post on Meta - this would allow discussion on whether or not it's a real problem. But it could also happen through a more formal process on /review - in this case, some amount of reputation would almost certainly be necessary to request a queue, and I'd even understand if this would be currently above my paygrade.
If done on a per-site basis, it would allow a particular Stack Exchange to specify its own problem tags, questions, and answers. On Stack Overflow, the homework tag is perhaps the most prominent example of a possible queue, but another is this post on how Accepted / Solved / Fixed is showing up in question titles. On English Language and Usage, there's the single word request. On Ask Ubuntu, a number have been highlighted in this post (which also includes a similar feature request). It goes on and on, and I'm sure more examples will come to light in answers and comments to this post. In all cases, an ad hoc review queue would allow the problem to be addressed in a recurring fashion, without resorting to banning a tag or title keyword - and hence, allowing legitimate uses to remain in place, assuming good quality reviews.
I do acknowledge, however, that good quality reviews cannot be assumed. Actually, part of my motivation for this is to help the rubber-stamping problem on the existing review queues - it's been highlighted that the thoughtful review is at a disadvantage, as the rubber stamp will quickly clear out the entire queue and receive all of the review points. Although disincentivizing the review process could help (for example, no Silver/Gold badge), just as problematic is that there just isn't enough content in the existing review queues (at least, for less than 3k users).
The ad hoc review queue would alleviate the scarcity - by establishing a higher threshold, there would be more items in a review queue at once, and there would also simply be more queues (and hence, items) to process. Hopefully, the lack of scarcity would encourage more thoughtful reviews, as there would always be more to look at.
In addition, the current per-queue badge levels could be replaced with a single set of aggregate badges over the whole review system (perhaps with higher counts), which would put even less pressure on snapping up items in any particular queue. They could also be augmented by a series of badges based on the number of review queues that you have contributed to, encouraging a wider understanding of the customs of a particular Stack Exchange.
It's been pointed out in the comments that this kind of feature has been requested a few times before - I'm totally open to closing this as a duplicate, should one of the other requests be implemented, or even declined. However, there are two things that make my proposal unique: first, with the threshold, it needn't become a permanent queue, forever cluttering the page with one or two posts. Second, it's not limited to tags only - any kind of queryable problem is available for the queue proposal.
Of course, the permanent queues could still be there for things that require more immediate action, like Suggested Edits (prevents others from editing) and First Posts (feedback to new users). If found to be widely applicable, an ad hoc queue could become a permanent queue at some point. But there would also be room to solve intermittent problems, that nevertheless need to be addressed by a larger portion of the community.