FBI struggles with data management

With a sniper killing people at random in the Washington, D.C., area in October and police pleading with the public for information, the FBI set up a phone center to receive tips and rolled out its computerized Rapid Start Information Management System to help sort them.

It sounds like an efficient, technology-driven process — until William Hooton described it in more detail in a Nov. 14 address to the Association for Information and Image Management.

Staff working the phones, including trainees from the FBI's academy in Quantico, Va., scribbled notes of their phone conversations on paper forms. Then, once every hour the forms were dumped in a box and delivered to the FBI's records management division. There, they were fed into scanners to be digitized, then added to the FBI's Rapid Start database, explained Hooton, assistant director of the FBI's records management division.

"Sounds like it's still a pretty manual process," said a technology vendor who listened as Hooton described the procedure. Scanning handwritten notes would yield a digital image, but not computer-searchable data, he said.

The FBI's use of records management technology during the sniper investigation illustrates the progress the bureau has made — and how far it still has to go — to bring its records management capabilities up to today's standards.

In March, Justice Department Inspector General Glenn Fine chastised the FBI and urged disciplinary action against several agents for mishandling records in the Oklahoma City bombing case. The discovery of thousands of misplaced records forced a month-long delay in the execution of Timothy McVeigh in 2001.

Fine blamed antiquated computer systems, but he blamed human error more for the FBI field offices' failure to turn investigation records over when they were ordered to do so.

Just weeks after Fine slammed the FBI, the bureau was rocked by another records fiasco — the "Phoenix memo." The FBI acknowledged that a memo an agent in Phoenix wrote questioning the number of Middle Eastern students attending U.S. flight schools never made it through the FBI's bureaucracy to senior officials in Washington, D.C.

The memo was written in July 2001, two months before the Sept. 11 terrorist attacks carried out by Middle Eastern men, some of whom had attended flight schools in the United States.

Even before the Phoenix memo became public, FBI officials conceded that they had a serious problem managing records. Their answer was to hire Hooton, an electronic records expert with experience at the Internal Revenue Service, the National Archives and in private industry. They also created a 1,000-person records management division.

Today, the division runs "10 production lines of scanners" and aims to convert 750,000 paper documents a day into digital records, Hooton said.

Once scanned, the electronic documents can be converted to electronic text by optical character recognition software, then stored in a database where they can be searched, mined and eventually made available to FBI field offices worldwide.

But converting such huge amounts of data to a more manageable electronic form is only one step toward solving the FBI's records problems, Hooton said.

After more than 90 years of collecting paper records, "we don't know what we have and what we don't have. We need to inventory our holdings," he said.

Other FBI managers have estimated that the agency has more than 1 billion records.

The process of digitizing, organizing and managing the FBI's records will never be complete, Hooton said, because there always will be a steady flow of new records.

Meanwhile, the FBI must tackle other thorny records issues, such as what constitutes an e-mail record, Hooton said. "No one has the answer now."

Similarly elusive is the answer to what constitutes a Web page record. And officials with the records management division need help deciding whether to destroy paper records once electronic copies have been made, he said.

Before the FBI finishes installing a modern computer infrastructure and records can be accessed by agents around the world, Hooton said he must tackle records security issues.

He said he favors maintaining centralized control over FBI records even if the records are physically housed in different locations.

"What the FBI is doing is way ahead of a lot of other federal agencies," said Jack Frost, a vice president at the electronic records management company TrueArc. Like the FBI, many agencies own huge archives of paper records, but few have the budgets or manpower to begin sorting and digitizing them.

Most federal agencies also are struggling with questions such as which electronic documents constitute official records, he said.

Once questions like that are answered, technology can help bring records under control, Frost said. Autocategorization software, for example, can sort records from other electronic documents and send them to databases for long-term preservation.

"Government is pretty much at [the] beginning stages of electronic records management," said another e-records specialist who heard Hooton's address but asked not to be identified. "The realization [is] only gradually dawning in many places that they need to do something about electronic records."

***

Taking inventory

William Hooton, assistant director of the FBI's records management division, said the agency doesn't know what records it has and doesn't have. He said he hopes to sort them into "three piles":

* One that is trash and should be destroyed. Hooton figures about half of the records fall into this pile.

* A pile that is "popular records," which should be converted to electronic form and retained because they have been retrieved at least once in the past five to 10 years.

* A pile of "unpopular records" that have not been used but must be kept. Those should only be made digital if they are requested, he said.