Agree 100% with Olaf that public opinion surveys aren't necessarily indicative of government policy (not hard to find things in the US where surveys show 80+ % support but that aren't what actually happens, e.g. on common sense gun control for example).
I personally, though, don't think people would generally look at something like SAI as "great, now we don't need to cut emissions" any more than people run red lights because they have airbags; I really don't think, if it winds up being used, that it looks like the world we were hoping for. I also find the argument that we shouldn't research it because there might be a moral hazard itself a deeply problematic position from a moral perspective, basically saying that rich people won't cut their emissions unless they watch enough poor people dying on TV.
But my question for Olaf on research is what line you think is hard to draw? Generally any report has come out saying some experimentation is ok and we shouldn't deploy or do "large" experiments, without ever coming up with firm definitions for what constitutes "large"... presumably that means that if there is good reason to conduct an experiment (because it meaningfully reduces relevant uncertainty, for example) and that experiment has negligible environmental impact, then it's ok, and obviously something that meaningfully reduces global mean temperature (say by 0.1C, so it's commensurate with interannual variability) would not be ok. (But we're so far from even being able to do the latter, and no-one is proposing it, and everyone would agree that that is deployment and not just research, that there's no reason to even debate that point). There are people who say "no outdoor experimentation, even if it releases 1 microgram", but those people presumably really don't want any research and are just using an experiment as an artificial boundary that they think they argue against, whereas it's harder to argue against modeling...
doug