Malintent detection: security, privacy, and the need for governance

The first three articles in this series of four  on malintent detection describe how these systems work, the ideas underlying their use, and an overview of applications and of their strenths and weaknesses.

Here, we examine the creepiness factor. People are accusing the authorities of attempting to invade our innermost thoughts with mind reading technology. Is this fair or accurate? Is our privacy being compromised in the interest of our security?

What is security?

According to Merriam Webster, security is freedom from danger. Taken literally, this definition would be unreasonable. No reasonable person can ever expect to be perfectly free from danger. Getting into a car carries a small risk of suffering an accident.  We accept this risk because taking it benefits us by getting us somewhere faster than if we were to walk. As Bruce Schneier often says, security is a trade off.

When accepting a risk in order to get a benefit, we also accept in principle the idea of a related trade off to reduce the risk without sacrificing the benefit. We won’t give up driving to reduce the risk of a car accident, but we’ve accepted wearing uncomfortable seat belts. We didn’t want to at first, but years of campaigns and mountains of data have convinced us.

Now when travelling by air we are subjected to intrusive scans and searches. The media have dramatically and graphically reported the risks and consequences of terrorism, so we accept the above intrusions in order to reduce this risk. Do these measures do more than just make us feel secure? It seems not; it seems they are merely security theater, showing us someone is doing something whether or not that something is effective.

Security Theater: Measures that make people feel more secure without doing anything to improve their security.

Real security should actually reduce risks or mitigate the adverse consequences of the risk. Given the costs to our dignity as well as to our wallets, is it too much to ask that security at least be effective?

We’ve shown that malintent detection can be very effective when used as part of a process, as one link in a security chain.  We should consider the trade off however. Are we in fact condoning an invasion of our privacy?

What is privacy?

Of course we are. There is no question that there is some intrusion going on here; there has to be.  The question is, is this intrusion reasonable and is there a possibility of the authorities abusing the trust we place in them? And what exactly are we losing as a consequence of this intrusion? What do we mean by privacy?

Daniel Solove leads the movement trying to define privacy. He points out that we disagree on even a basic definition of privacy. Many plausible definitions feel right at first but some proposals prove too broad while others too narrow.

For example we might define privacy as controlling one’s own personal space. But if someone shoves us in a crowded bus, do we think our privacy has been violated? Probably not.  The definition is too broad.  Alternately, we might define privacy as intimacy. But if someone steels your social security number, you’d certainly think your privacy has been invaded but you would not feel your intimacy has been violated. The definition is too narrow.

Daniel Solove's family of related privacy concepts

Instead of trying to define privacy as a single idea, Solove proposes we consider privacy as a family of related concepts.

We tend to think of privacy as the right to hide something and when we accept a search, we give the authorities the right to check if we are hiding something. If we have nothing to hide, we ought to have no complaints.

The “nothing to hide” argument involves the “surveillance” concept from the family of related privacy concepts. It is reminiscent of George Orwell and his dystopian novel “1984″. We would not accept this argument if someone were to search our homes or listen to our private conversations without a warrant. On the other hand our demand for real security makes this a powerful argument for searches and scans at airports.

But as Solove argues, there is more to privacy than the right to hide something. There’s a non Orwellian aspect to our privacy that has nothing to do with surveillance. It’s a Kafkaesque quality of intrusion that leaves us feeling helpless.

Orwell and Kafka illustrate two members of the family of concepts in Daniel Solove’s model of privacy.

Franz Kafka‘s “The Trial” is about a perfectly ordinary man who is arrested and put on trial without ever meeting his prosecutor, his accuser, and even without ever being told of what he is being accused. He wakes up one day and finds his life turned upside down.

Many people have experienced something like this. They show up at an airport and are refused boarding because they are on the no-fly list.  Even the late senator Edward Kennedy was denied boarding on a flight to Washington because his name appeared on the no-fly list. Investigation showed that the name did not refer to him but to an Irish terror suspect also named T. Kennedy. If a United States Senator can’t easily fix a mistake like this, what chance do ordinary people have?

Where body scanners are more Orwellian, the no-fly list is more Kafkaesque. We grudgingly accept the former (in principle, but effective security would be better) but we strongly oppose the latter. But where do malintent detection systems fit in this scheme?

It seems a short step from mal-intent detection to lie detection. Gad Saad of Concordia University wrote about how many times a day we lie. Do we want buzzers ringing every time we tell a little white lie?

It’s not mind-reading

But it’s not really lie detecting and it’s certainly not mind reading. Malintent detection looks for responses to familiar stimulus. It doesn’t dig into our minds, it reads the evidence that jumps out of our bodies. There is indeed something Orwellian about this but this should not worry us: we accept surveillance in the proper context.

What we ought to fear are the Kafkaesque attacks on our privacy, the no-fly list type of abuse to which the authorites might subject us for their convenience. These systems must not be used as labeling machines forever branding us as guilty. They should be used as tools that merely hint that someone should be examined a little more closely before being let through.

The need for governance

Henry Kissenger is a great fan of Cardinal Richelieu and in his book Diplomacy he credits him with inventing the modern nation state. He was a cunning politician who preferred to rig trial with well chosen evidence rather than rely on clumsy assissination plots.

Armand Jean du Plessis, duc de Richelieu

“Give me six lines” said Richelieu “written by the most honorable of men and I will find an excuse in them to hang him.” It is this sort of abuse we must prevent.

We need to be protected and we must allow our agencies to use the best tools and their best judgment in doing that, but we cannot allow authorities to abuse the power they derive from these tools. Malintent detection systems are powerful tools and their use must be controlled in order to benefit us, not to harm us.  The technology is here and it will be used. It’s up to us to demand it be used properly.

Copyright 2011, Vincent Poirier

Disclaimer: I am helping promote a malintent technology system developer in Japan, so please be aware of the possible conflict of interest.

Explore posts in the same categories: Governance, Information Security, Information Technology Management, Opinion