Safe surveillance?

In my last post I made the reluctant argument that we need to come to terms with the need for surveillance. Given the increasing danger that one person or a small group can cause almost unimaginable horror, we have to be able to prevent it.

Is it possible to have surveillance without tyranny? I have a glimmer of an idea.

The problem is basically one of scale. If there is a file on every US citizen, then many thousands of people will have to be involved in gathering the information, collating it, putting it into patterns, making decisions. There are so many points at which that information may be misused or hacked that it’s basically impossible to prevent. That path leads to nothing but bad results.

But the technology community has begun to develop an answer. It’s called “homomorphic encryption”. It makes it possible to process encrypted data without decrypting it first. The Hacker Lexicon says:

“A homomorphically encrypted financial database stored in the cloud would allow users to ask how much money an employee earned in the second quarter of 2013. But it would accept an encrypted employee name and output an encrypted answer, avoiding the privacy problems that usually plague online services that deal with such sensitive data.”

In other words, the many thousands of people who touch the data could gather and analyze it without ever being able to see people’s names or identifying information. There might be “files”, but no one looking at those files  could tell whom they referred to. They could sift it for patterns that might indicate terrorist plotting, but they would not know to whom those patterns pointed.

That might make it feasible to cut down actual intrusions into personal privacy far enough to bring to bear the traditional mechanism of protection: judicial review. When law enforcement officials found threats in the data, they could ask a judge for the right to decrypt it. It would be like asking for a wiretap. It’s a violation of privacy, but one hedged with enough protections that we have agreed it is worth it to reduce crime. 

Homomorphic encryption is technically very difficult and still quite slow, but it has been improving rapidly and is getting close to feasibility. If this could work, it could reconcile defense against the growing threat of lone-wolf attacks with the rights of personal privacy and dissent.