To be honest, there is less to this story than meets the eye, though you cannot tell that from the headline and introduction.

An internal FBI audit has found that the bureau potentially violated the law or agency rules more than 1,000 times while collecting data about domestic phone calls, e-mails and financial transactions in recent years, far more than was documented in a Justice Department report in March that ignited bipartisan congressional criticism.

The new audit covers just 10 percent of the bureau's national security investigations since 2002, and so the mistakes in the FBI's domestic surveillance efforts probably number several thousand, bureau officials said in interviews. The earlier report found 22 violations in a much smaller sampling.

Sounds bad -- until you read the next paragraph.

The vast majority of the new violations were instances in which telephone companies and Internet providers gave agents phone and e-mail records the agents did not request and were not authorized to collect. The agents retained the information anyway in their files, which mostly concerned suspected terrorist or espionage activities.

So the reality here is not that the FBI acted inappropriately, but that they received more information than they initially requested. Actual instances of inappropriate requests were a mere handful -- about two dozen.

Now why would the FBI keep the additional information received and how was it justified? Well, the Post explains that here.

Of the more than 1,000 violations uncovered by the new audit, about 700 involved telephone companies and other communications firms providing information that exceeded what the FBI's national security letters had sought. But rather than destroying the unsolicited data, agents in some instances issued new National Security Letters to ensure that they could keep the mistakenly provided information. Officials cited as an example the retention of an extra month's phone records, beyond the period specified by the agents.

Case agents are now told that they must identify mistakenly produced information and isolate it from investigative files. "Human errors will inevitably occur with third parties, but we now have a clear plan with clear lines of responsibility to ensure errant information that is mistakenly produced will be caught as it is produced and before it is added to any FBI database," Caproni said.

So what we have here is a situation in which the procedures were unclear and so agents acted to bring the information within the scope of the law after it had been erroneously given by someone else. Really, can you blame them? Would you want to be the FBI agent called to testify before Congress following a terrorist attack and have to explain why you destroyed a key piece of evidence that could have nipped the plot in the bud? I know I wouldn't -- and no amount of good faith in having done so would ever allow one to be absolved of doing so in the eyes of the American public. As it is, I expect that some of those records that have been isolated under current procedures could be the source of much contention if they ever prove to hold a "smoking gun" in a future terrorism case.

Oh, and the actual rate of violations?

FBI officials said the audit found no evidence to date that any agent knowingly or willingly violated the laws or that supervisors encouraged such violations. The Justice Department's report estimated that agents made errors about 4 percent of the time and that third parties made mistakes about 3 percent of the time, they said. The FBI's audit, they noted, found a slightly higher error rate for agents -- about 5 percent -- and a substantially higher rate of third-party errors -- about 10 percent.

Higher than we would like, but not an unreasonable rate given that we are dealing with a period that included the earliest anti-terrorism investigations with the new tools given investigators in the wake of 9/11. I wish we had a breakdown of these errors over time, to see if the rate has decreased from the earliest days.

Posted by: Greg at 12:49 AM | No Comments | Add Comment
Post contains 679 words, total size 4 kb.