While I dislike mass surveillance as much as the next guy, the OP's particular argument against it strikes me as myopic. The risks which PRISM purportedly addresses do not cap out at the 3000 deaths by terrorism in 2001. One nuclear device detonated in a major city (let's say in a RORO container awaiting inspection in New York's harbor) could kill millions. Does inclusion of that number in the math make PRISM rational?