Skeptical • November 6, 2013 6:03 PM

65535 –

Would it not be as much as an error to conclude from the Boston Bombing that current security is a total failure as it would be to conclude from an airplane crash that current flight safety efforts are a total failure?

No security is going to be perfect. No system of safety is going to be perfect. We can work on reducing probabilities, but they’re unlikely to ever get to zero. So while we learn from every incident, and improve with every incident, we’ll always need to prepare for future incidents.

Given the difficulty of the task, my own view is that the record of prevention is impressive.

Matt –

I’d speculate that such a tactic (using one attack as a diversion to stage an attack elsewhere) is an anticipated, even expected, contingency. When something happens the response always entails more than sending personnel to the reported scene. Immediately after the Boston Marathon incident security was increased not just across Boston but across the United States, especially in cities most sensitive to the possibility of terrorism, no?

Clive –

It’s been reported that new buildings in Pakistan are now sometimes designed using software that can simulate the effects of explosive devices of various types, in order to create buildings that mitigate the effects of such devices.

I would say (I know I’m likely in the minority here) that better surveillance has a role to play as well. Higher quality public surveillance cameras with broader coverage, software that can recognize possibly suspicious actions (leaving a backpack unattended) and direct attention, faster facial recognition, better scanning technology for concealed devices and weapons, and, yes, signals intelligence that can be used to thread together what might be nearly disparate cells or to paint the past of a hitherto unknown terrorist.

As bad as the Boston bombing was, it’s easy to envision even worse scenarios had the two perpetrators been determined and able to travel to New York City and launch another attack.

In other words, containing an attack means planning not just to mitigate any additional explosive devices, but also to find and “defuse” the perpetrators of the attack. This would be important not simply from the vantage of limiting further physical damage, but also from the vantage of limiting any further psychological or economic disruption from an attack. And that in turn is an important part of discouraging future attacks.

I completely understand the civil liberty concerns with such technology, but we ought to be able to design institutions and procedures in such a way as to enable the appropriate development and use of that technology while protecting against its abuse. And there are benefits beyond preventing and mitigating terrorist attacks, such as detecting reckless or intoxicated drivers, persons in medical distress, crimes in progress, or wanted violent criminals. Properly limited and supervised, technology like that can reduce not just low probability/high impact events, but also the more frequent causes of death and suffering that occur every day.

Skeptical • November 7, 2013 6:21 AM

Brandioch –

Good point, but it depends on how good the software’s heuristics are and how large a response any given software signal will bring, no? If it signals every time a person simply rests a backpack or bag on the ground, then not very good and little response. If it signals whenever someone throws away a small bit of trash, not very good and little response. But if it picks out the guy who walks to a point in a crowd, dumps a large bag, and then rapidly walks away… well that’s a good flag, and should bring some immediate response to verify the flag. By immediate response I mean at first something as simple as a human being taking a closer look at the footage flagged by the software; it can be escalated or de-escalated from there.

And with machine learning approaches, the software can improve every time it’s shown an instance of an actual bombing or attempt (or cardiac arrest, or mugging, or drunk driving, etc).

Re SWAT teams and curfews, this type of disruption is minimized if you’re able to identify the perpetrators and quickly find and fix them. But, if you don’t have that capability, and you’re grappling with an ongoing terrorist operation, then a voluntary curfew and lots of requests to conduct searches (99% of homeowners will be grateful for the check) will be used.

My view is that we’re not heading towards a police state, but rather to a time when more and more public behavior will be verifiable. And that means we can tell with more accuracy whether a police officer used excessive force, improperly checked a person’s records (in the US, checks of national databases by law enforcement are already audited regularly by the FBI), etc. It’ll mean very careful regulation, transparency, and control over the technology enabling all that, but ultimately this is something that should enhance legal rights, not diminish them.