This is a long term solution only and is only applicable to such things as coursework and where the science is well known with predictable results. It won't work, most likely for doctoral level research experimentation.
But, it is possible, though costly and time consuming, to build realistic simulators that let a student learn both lab technique (perhaps imperfectly) and how the science works. A truly realistic one would introduce random effects as occur in real experiments, of course. And, such things will only take you so far.
But, at the moment, the need for such things has only been obvious for a relatively short time; too short to expect that such simulations are readily available. Some do exist, I'd expect, but not with the necessary sophistication to foster insight in students. There are probably some adequate simulations available in physics and chemistry, for example. Biology and pharmacology are harder. But these are, again, helpful for coursework, but likely not for serious research.
And, the people who know best how such things work aren't the same people capable of creating such simulations themselves. It is, itself a team process with a number of different skills contributing.
I suspect that the need for this will continue into the future, whether we finally find a cure for COVID or not since the disruption has caused people to rethink how courses are delivered. We might not ever go back to the pre-covid world.
Some inducements are probably necessary to get people to build and refine such simulations. And some inducements to get professors to consider using what is currently available in the short term.
Other solutions are possible at the fringe, though also difficult to use. Labs where only a single person works are probably too expensive in most cases. Housing a group of experimenters together so that they become something like a "family" is also probably too radical to adopt, though some sports leagues do this so that the game can go on.
At the moment, we need to "think outside the box" but also plan for a possibly disrupted future. Personally, I doubt that Omicron is the last of the variants and there is no guarantee that the next one will be of less concern.
It might require some push from university administration to get some collaboration between various professors in the experimental sciences and those in computer science. Some of the simulations would be good upper level student projects in CS.