Surely what MIRI would ideally like to do is to find a way of making intelligence not "emergent", so that it's easier to make something intelligent that behaves predictably enough to be classified as Friendly.
find a way of making intelligence not "emergent"
I don't believe that MIRI has been consciously paying attention to thwarting undesirable emergence, given that EY refuses to acknowledge it as a real phenomenon.
http://singularity.org/blog/2013/01/30/we-are-now-the-machine-intelligence-research-institute-miri/
As Risto Saarelma pointed out on IRC, "Volcano Lair Doom Institute" would have been cooler, but this is pretty good too. As the word "Singularity" has pretty much lost its meaning, it's better to have a name that doesn't give a new person all kinds of weird initial associations as their first impression. And "Machine Intelligence Research Institute" is appropriately descriptive while still being general enough.