Man, this is exactly why I dislike this paper. Somebody reads abstract and conclusion section, sees those probabilities, and then runs with them, possibly forming an entire world view around that. But there are such massive asterisks that should be attached to these numbers which are completely absent in that kind of presentation.
The ~50% probabilities from the paper are not fundamentally more founded in reality than "Did another planet in our galaxy develop intelligent life yet that attempted to communicate? I don't know, so let's assume 50% either way. Therefore, we are alone with 50% probability." Should that number be 50%? Or 10%? Or 99%? You're choosing some prior there. So does this paper, and the priors chosen gravely affect the result. Of course, the paper has more intricate math and more solid reasons for their priors, but fundamentally it is still just assigning almost arbitrary probability distributions to parameters that we have no grasp on. Specifically (see my other comment), the numbers that they get primarily depend on the choice of prior distribution for the abiogenesis rate, which overshadows all other parameters. It has such massive error bars (>200 orders of magnitude by their own account) that arriving at a concrete probability (in the way that you understood it when reading the paper) is somewhere between misleading and a lie, because it's more or less just a projection of that prior into a single number.
In other words, I'm convinced they could have easily arrived at probabilities like 0.1% or 99.9% by reasonably varying their prior distribution for that parameter alone.