This debate seems like a manifestation of a problem with governments: They think they can legislate anything they want. Need access to some communications - green light for massive data collection. Some of it is encrypted - just mandate a back door. School shooting - new gun laws. Any problem activity - we'll just make it illegal. Something not getting done - we'll just mandate someone take care of it. They really do…
In other words, it's a naïve approach.
For example, take the Digital copyright problem in the music industry. Almost anyone would agree musicians should get paid; so legislators feel they have to 'solve it'. And indeed legislators can 'solve it' using just that naive approach: make it illegal, put a huge fine to it. So when you download a music, you will consider the risk of getting caught. The issue then is obvious: if you only legislate, the chance of getting caught will be very low, so most people won't care. But if you set the fine/punishment high enough, it's pretty much a given people will start being careful. But how much is enough? If you have a 1 in a million chance of getting caught this has to be astronomically large: if people simply thought about expected value (we don't exactly, just to illustrate), you'd need to charge millions per infringement. It's a totally unreasonable punishment, basically a terrible system. This is (an example of) why symptomatic legislation/treatment often doesn't work: it's grossly ineffective/inefficient, it's usually best (but way more complex/demanding) to address the root causes.
So the best solution to terrorism/criminal activity is obviously not a surveillance state, but unfortunately it works to an extent. I feel legislators don't have at their disposal enough tools to handle the complexity of better solutions. It would be good if we could involve academia/industry way sooner in direct collaboration with legislators, reaching effective solutions -- I feel science is indispensable in finding them for most complex cases.