Governments are obsessed with creating (or restoring, depending on who you talk to) a power imbalance they say is necessary to maintain order and prevent serious crime and terrorism.
I and many other technologists argue to the contrary - that blanket surveillance will have little long-term impact on seriously organised criminal and terrorist activity as perpetrators will adapt quickly to evade today's proposed imperfect monitoring systems as technology continues to evolve quickly.
At most such technology will trap mid and low-level criminals, giving a short-term advantage that will soon be lost as even petty thieves learn how to hide their online trail more efficiently.
Whilst reducing crime of any description is undoubtedly a good thing, this must be balanced against the cost and risk to all non-criminals in society who face having even more of their secrets held by state agencies and other third parties without their consent.
Additionally, such monitoring carries a significant cost.
Not just the monetary cost in siphoning off, storing, filtering and retrieving large quantities of data - but a cost to technological progress. Internet service providers may shun network upgrades because of the added complexity of accommodating the surveillance regime, denying customers increased bandwidth and other benefits of the latest technology.
Additionally - or so a network engineer at a very large mobile phone network once told me - network changes required to meet today's data retention laws made the system, in his opinion at least, more vulnerable to failure because all transactions had to be routed through one of a few data collection points.
Police in the UK still don't carry guns on routine patrols - why? Because the risk outweighs the benefits. Society is generally better off with a softer balance of power - consensus policing - and not carrying a firearm is a powerful reminder to the public.
The benefits of a more consensual approach to policing is that the public are more likely to do their bit to help the police in their duty; contrasted with more militaristic approaches, which pit the public - even the law abiding public - against the police, whom they often live in fear of.
I believe analogies can be drawn with policing the internet. I'm afraid of snoopers taking a snippet of my data out of context or misidentifying someone else's transaction as originating from me. I'm afraid of a large mountain of my personal data leaking, leaving me vulnerable to identity theft.
I'm afraid of police creeping around in bushes watching ordinary citizens go about their lives - because this, quite frankly, just freaks me out.
Of course the state must be involved in some way; a free-for-all leaves the weak unprotected.
But the level and manner of involvement I have in mind usually contrasts strongly with what governments around the world are pushing for.
I believe the internet should be policed to a large extent via the front door, not by creeping around the back or hiding in bushes with the digital equivalent of a long-lens camera and parabolic microphone.
The Home Office often cites the hunt for dangerous paedophiles as justification for blanket surveillance, playing to the public's fears.
A group of vigilantes recently showed us all that progress in the fight against paedophiles can be made without snooping around behind the scenery planting bugs in the very fabric of the network.
Police condemned the action of vigilantes as potentially illegal itself, but this perhaps says more about police wanting to maintain an illusion of control, or says something about the contradictory state of current privacy laws which are seen by some as limiting police operationally whilst allowing the state to watch us all via our mobile phone activity, etc.
If we could find some way for police to use the internet via the front door, connecting via an ISP to inhabit the places people hang out online - in a similar way to the mix of visible and plain clothes patrols inhabiting the streets; then this surely will be more proportionate and more sustainable than relying on blanket surveillance.