Welcome to Cato Daily Podcast for Friday January 5th. I'm your host on

Nastasia Glova. Today our guest is director of information policy studies

Jim Harper who along with Jeff Jonas of IBM, authored the Cato Policy Analysis,

effective counterterrorism and the limited role of predictive data mining.

In the report, Jim explains his opposition to the use of data mining techniques

for identifying potential terrorists, which is the topic of today's podcast.

How much information was available about the attackers prior to September 11?

There was quite a good deal of information available, because it was easy, frankly, for

newspapers, in fact, to say nothing of investigative authorities, to track down who had

been where, who had done what, and there was a lot of information about the

911 attackers and how they had interacted, the phone numbers they shared, the frequent flyer

numbers they shared, the addresses they shared, and so on and so forth.

So had we been looking for the two 911 attackers that were known to be in the United States

to authorities, we would have probably been able to wrap up the 9-1-1-1 attack, at least this is the

conclusion of the 9-1-1-commission.

How does data mining fit into this then?

Is it useful in discovering and identifying threats? Well, data mining came into vogue shortly after 911 as a response to 911, or perhaps better put a reaction to 911,

because it's probably not a well calibrated response.

First let me go into what data mining is.

You can look at data analytics in two ways really.

There's link analysis that's following one piece of data to another.

So if you know the phone number that's been used for crime,

...