Encryption will make a signal look like noise, but first you have to be able to separate that source from the host star, and I think even doing that is difficult. I think that's the more fundamental problem.
Could be but first you have to find a candidate star. And with all radom noise where would you start
I think you're kind of missing the point. According to the Drake equation and what we know about exoplanets now, you start by pointing to any star, any at all. The problem is that you won't be able to distinguish signal from the planet(s) from signal from the star.