By Jesse Marx
By Chris Parker
By Jake Rossen
By Jesse Marx
By Michelle LeBow
By Alleen Brown
By Maggie LaMaack
By CP Staff
As he gives the nickel tour of his workplace, Tom Kooy stops now and then to provide a little narration. The office is large--about 30,000 square feet--and oddly empty. Kooy acknowledges that it's more space than the dozen or so staffers who work here need. But the state got a good deal on the space. Besides, he adds, it's really just a stopgap measure while work is finished on the new Bureau of Criminal Apprehension headquarters in St. Paul. So, for the time being, Kooy and his colleagues are charting the future of the state's criminal justice system from Two Pine Drive, a colorless and entirely unremarkable office park in the St. Paul suburb of Arden Hills. As we pass through a maze of cubicles, the persistent hum of computers is audible in the background, overlaid with the muffled conversation of worker bees talking to each other over cube walls. We pop into a vacant conference room, where a dry-erase board is dappled with a dizzying array of acronyms--MNCIS, MRAP, POR, SSS--and chicken-scratch diagrams. "This is the war room. It's where our technical people meet," Kooy says.
Finally, Kooy stops in a hallway and gestures toward a locked door. Behind the door is the heart--or, perhaps more accurate, the brains--of the operation: about $6 million worth of computer hardware and software that will serve as the hub for a complex and extremely ambitious computerized crime-fighting program called CriMNet. The goal of CriMNet, Kooy explains, is relatively simple: to allow for the instantaneous exchange of information among criminal justice agencies. But carrying out that goal will be anything but simple. There are more than 1,100 such entities in the state--police departments large and small, sheriffs, county attorneys, clerks of courts, probation offices, and so on. Eventually, all those agencies are expected to tap into the system, sharing information with one another to a degree that until very recently would have been unimaginable.
Theoretically, Kooy explains, CriMNet will function much like an Internet search engine. The users will punch an individual's name into a field on the computers. Then the CriMNet hub will transmit the query to the appropriate agencies, and a digital dossier will pop up on the user's screen. If the subject of the query has been arrested before, the display will include a number of standard elements, including a mug shot from the Minnesota Repository of Arrest Photos (MRAP) and a comprehensive criminal history. The criminal histories will be far more extensive than the ones currently available from the BCA. (As a rule, BCA criminal histories include only felonies and gross misdemeanors; and they are notoriously incomplete.)
CriMNet will also draw on a host of other state and local government databases, including motor vehicle records, information on outstanding warrants and restraining orders, and probation status. It will even provide users access to details about an individual's most casual contacts with police. Employing a database called the Multiple Jurisdiction Network Organization (MJNO), CriMNet users will be able to find out if you were ever pulled over for a cracked windshield or questioned by cops at a loud party. There are some six million records on the MJNO, even though fewer than a third of the state's 500-plus police agencies currently participate. Under CriMNet, every department in the state will be enrolled.
"I don't think any other state has even tried to do anything on this scale," says Kooy, who serves as the deputy director of the CriMNet project. "Basically, we're trying to put together a holistic, automated system so that all the different agencies can share their information." Technically, it is an extremely complex project that will require custom-built software, including specialized adapters to allow communication between different types of databases. And then there are the logistical and political hurdles. CriMNet, Kooy is fond of saying, is not a state project, so much as "a statewide project." And its ultimate success is predicated in large part on local government entities investing the time, effort, and money to make it work.
These are heady times for CriMNet's backers. On February 3, CriMNet and a related project called MNCIS (for Minnesota Court Information Systems) made their debut in Carver County. A lot is riding on the success of the pilot program. "We're about to turn a very important corner," says Kooy. "And I think most people need to see something tangible, because it's difficult for nontechnical people to understand until we build and deliver it."
CriMNet has enjoyed near-unanimous support among lawmakers and other power players in the state, ranging from Charlie Weaver (the Pawlenty administration's chief of staff) to the late Sen. Paul Wellstone. Newspaper stories have celebrated it. Corporate interests, including the Target Corporation and a coalition of business leaders known as the Minnesota Business Partnership, have been heavily involved in promoting and marketing the program. CriMNet boosters have sought to position the system not only as a lifesaver to the cop in the field but also as a template for the federal government's intelligence-sharing operations.
Despite all the public expressions of support, there are plenty of reasons to be skeptical about CriMNet. And there is a growing, if still small, cadre of critics waiting to be heard. No one has publicly objected to the stated goal of the program: improving the flow of information within the criminal justice system or, as the CriMNet motto has it, "putting the right information in the hands of the right people at the right time." Almost everyone who knows anything about cops and courts admits that incomplete or inaccessible record systems have been a chronic problem in Minnesota.
But as the old saw goes, the devil is in the details. And if CriMNet is rich with one thing, it's details. Questions about cost of the program have taken on a new heft in light of the state's budget crisis. CriMNet staffers have at various times calculated the total program cost at $100 million to $260 million over the next six years. The truth is, no one knows what the final tab will be. Republican state representative Phil Krinkie, one of CriMNet's most vocal critics, contends that the $260 million figure is nothing but low-balling. "They're obviously trying to keep those numbers down because they don't want to scare people into thinking this is going to be a billion-dollar system," Krinkie says.
And then there are other, even more robust veins of criticism, ones based in concerns about civil liberties and fairness. If CriMNet works as designed, vast reams of information held by the 1,100-plus criminal justice agencies in the state will suddenly become retrievable with a click of the mouse. That, of course, creates opportunities for abuse. Will some elected officials (county attorneys and sheriffs, say) be tempted to tap into the databases to find dirt on a political opponent? And what will happen if the criminal background checks made possible by CriMNet--so much more comprehensive than the ones that exist today--wind up in private hands? Or on the internet?
Robert Sykora, an attorney who works as information systems director for the state's Board of Public Defense and serves as co-chair of the Minnesota Criminal and Juvenile Justice Information Task Force, is worried by such scenarios. "I think CriMNet has the potential to do some good things. I don't challenge that," Sykora says. "But my concern is the negative consequences. What's coming along in the Trojan Horse?"
If CriMNet had an official mascot, it would be Donald Blom. No name has been invoked more often or more passionately by supporters to demonstrate the need for the project. In 1999 Blom, a then-51-year-old Richfield resident, was arrested in connection with one of the more notorious crimes in recent state history, the kidnapping and murder of a teenage convenience-store clerk named Katie Poirier. The abduction was recorded on store video. Poirier was an attractive and sympathetic victim. Her body was never found. So, not surprisingly, the story appeared on TV continuously. When the cops finally caught up with Blom and details about his troubling past emerged, a media frenzy ensued.
It turned out that Blom had an extensive history of convictions for sexual assaults and other crimes. But he was clever. Over the years, he used aliases and claimed different birthdays. Because of gaps in information sharing and identification- verification procedures (mostly related to poor fingerprint matching) among the various agencies involved, Blom's full record was not known. Not to his employer, a veterans' home where he had passed a BCA criminal history check; not to the judges who sentenced him in the previous cases; and, most critical, not to the investigators who were desperately hunting for suspects in the days following Poirier's disappearance.
"In the case of Blom, the counties and police departments all had information on him but they had no way of sharing," recalls Public safety commissioner Rich Stanek, a former state representative and Minneapolis police lieutenant who is one of CriMNet's most ardent boosters. "If CriMNet was in place then, right after Katie Poirier was abducted, they could have gone into all the databases, and searched for sex offenders--by m.o., by geography, by motor vehicle records. But we couldn't. That was Minnesota's dirty little secret."
Among cops and prosecutors, of course, it wasn't a secret at all. For the better part of the decade, both the state and Hennepin County had been working on so-called integration projects, which were aimed at improving the flow of information among various agencies. The Blom case, and the weaknesses it exposed, had a galvanizing effect on state lawmakers. After hearing testimony from the mothers of numerous high-profile victims (Poirier, Jacob Wetterling, Cally Jo Larson), the legislature quickly cobbled together and enacted a $41 million anti-crime package. Dubbed "Katie's Law," the new measure effectively transformed the as-yet-unnamed CriMNet program from an obscure little Hennepin County project into a much broader, much more ambitious statewide initiative.
To Stanek and the project's many other backers, it was a no-brainer. After all, as Stanek points out, the private sector long ago realized the benefit of computerized inventory tracking; shouldn't cops and prosecutors have such state-of-the-art technology on their side too? "If I go into a Target in Maple Grove, and they don't have the black polka dot socks I want, they can tell me which of their 100 stores in Minnesota have them," says Stanek. "Well, I couldn't give you that same type of information about Level 3 sex offenders, the worst of the worst."
Stanek's use of the Target analogy is fitting, if for no other reason than that the company has taken a keen interest in the development of CriMNet. Target's involvement with criminal justice integration issues in Minnesota dates back to the mid-'90s. According to Nate Garvis, the company's director of government relations, there were two major triggers: a crime wave in Minneapolis (which infamously led the New York Times to dub the city "Murderapolis") and a Star Tribune investigative series about recidivist offenders called "Catch and Release." The Strib story told of lawbreakers who had received light sentences because the full extent of their criminal records was not known to prosecutors and sentencing judges.
"My CEO asked me to look into it, and asked, 'Isn't there something we can do?'" Garvis recalls. Since then, the company has assumed a quasi-official role in the push to build CriMNet. It has supplied technical advice, lobbied legislators, and even hired the public relations firm Himle Horner to write newspaper opinion pieces. Target's motives? Entirely altruistic, says Garvis. "This is just a broadening of our corporate philanthropy," he says. "There's no parochial interest for Target in doing something like this. Our interest is that it's good for our communities, and that we'll all gain by having safer communities. There's no business rationale behind our activities here."
That's open to debate. For big employers like Target, CriMNet offers a potentially invaluable resource: access to faster and much more comprehensive criminal background checks. For retailers, knowing whether or not a prospective employee has a history of theft or other crimes is surely useful information. Criminal history checks have become a routine procedure in many sectors of the economy now. And with reason. In a nationwide study conducted by the University of Florida in 2001, employee theft was found to be the single largest cause of "inventory shrinkage" among major retailers. The study, which surveyed some 120 businesses, estimated annual losses from employee theft at approximately $15 billion. (Shoplifters, for their part, caused some $10 billion in losses.)
But retailers and other big employers are hardly the only private-sector entities that stand to benefit from better access to more detailed and complete criminal histories. Landlords, for instance, routinely contract with screening companies to check into the backgrounds of prospective tenants. Such companies draw on many different databases, including credit bureau reports, and develop their own databases of problem renters and kindred undesirables. But the BCA background check is the best single source of information for agencies in Minnesota, according to Lee Michaelson, vice president of the Twin Cities-based Rental Research Services. Access to an improved and more complete criminal history, he says, would certainly benefit his business.
Target's Garvis point out that it has not been decided whether anyone outside of the criminal justice community will have access to even limited amounts of CriMNet data, such as conviction information. In fact, the Criminal and Juvenile Justice Information Task Force, a group consisting chiefly of government bureaucrats who are making recommendations on CriMNet, has urged that CriMNet be exempted from the state's open-records laws until 2004 while they untangle the nettlesome issue of public access.
That, in its own right, is problematic, says Don Gemberling. The state's leading authority on the Minnesota Data Practices Act, Gemberling serves as the director of the Information Policy Analysis Division at the Department of Administration. "I know from bitter experience that if the detailed requirements of the Data Practices Act are not integrated into computerized systems at the design level, then guess what? You wake up one day and you're told that it can't be made compliant because it's going to be too damn expensive," Gemberling says. "If I've heard that once, I've heard it 40 times."
Privacy advocate Rich Neumeister shares Gemberling's concern. One of four citizens appointed to serve on a CriMNet advisory panel, Neumeister worries that fair information principles, the foundation of the Data Practices Act, are being given short shrift. "This technology raises a lot of questions about the relationship between citizens and state authority. Policymakers need to ask those questions," Neumeister observes. "But it's not really happening, because the techies are running the show." Take MRAP, the state's newly developed digitalized mug shot database. With the coming generation of facial recognition software, Neumeister points out, MRAP has the potential to become a cornerstone of sophisticated surveillance operations. But there has been little discussion about whether that ought to be limited or prohibited. And, Neumeister says, the issue of access to criminal histories remains murky.
As it stands now, you can obtain criminal histories by two means: making a written request to the BCA and paying a fee, or going to the BCA headquarters in St. Paul and entering search queries into the terminals there. Technically, all felony and gross misdemeanor convictions less than 15 years old are public. Arrest data, on the other hand, is deemed private--a rule rooted in the principle that people should not have to bear the stigma of an offense for which they have not been convicted. (Interestingly, that same data is public data if you get it from the local police department instead of the BCA. It is a bureaucratic obfuscation put in place to serve privacy interests.) But there is a huge loophole. If the subject signs what is known as an informed consent form, then private arrest data can be released. And landlords and employers routinely require that prospective tenants and employees sign the consent form. Of the more than 116,000 background checks conducted by the BCA last year, 98,850 were performed with the consent of the subject.
So what will the consequences be for people looking for work or housing when the more expansive and detailed criminal histories found on CriMNet become available? "I'm afraid we are going to create a class of people who can never be employed and never find housing because every peccadillo and mistake will follow them for as long as they live,"posits the Board of Public Defense's Sykora. He is especially alarmed by the prospect that criminal records will become available on the Web. "There is a real and palpable difference between information that you can only get by standing in line in a courthouse and information that is available to anyone on earth with a web browser," he adds. "So we've got a huge public policy decision we've got to make before we transfer this to the Web."
Sykora's biggest concern, however, can be boiled down to a single word: accuracy. Minnesota court records, Sykora say, are often incomplete or inaccurate. As it is now, such errors or omissions can carry major consequences. Take this common scenario. A defendant is charged with a felony but pleads down to a gross misdemeanor. Often, he says, that gross misdemeanor conviction will appear in some court record as a felony. "The sentencing guidelines will assign certain points based on your criminal history. So if you get in trouble again, and you walk in with a felony point, and nobody catches that error, you'll end up going to jail longer than you should," Sykora says. "Our lawyers find errors like this all the time."
Mark Anderson, who works in the public defenders' appellate division, agrees. In one of his recent cases, Anderson says, he represented a client who had previously faced a charge of fleeing police. Ultimately, the man pleaded guilty to a much lesser offense, making an improper turn. But it was the fleeing-police charge that wound up on his criminal history: The clerical error effectively added 10 months to his sentence. As Anderson sees it, the problem boils down to issues of overwork and inadequate quality control among low-paid data entry clerks.
But with CriMNet, Sykora and Anderson contend, the consequences of bad information will be amplified. Information that was hard to get--sometimes existing only in paper form in the back of a filing cabinet--will now be universally available. Or, as Sykora puts it, data once distributed "by the teaspoon is now going out by a firehose."
With an eye toward assessing the problem, the Criminal and Juvenile Justice Policy Group--CriMNet's governing board--has requested that the legislative auditor undertake a study in which cases are tracked as they wend their way through the court system. Such an audit, says Sykora, would help define the error rate. In principle, that seems like a good first step. But it is only a recommendation. With tight budgets, there is scant guarantee the auditor will mount such a study.
The Policy Group, meanwhile, has also been looking at creating a mechanism whereby people could challenge the accuracy of information on CriMNet. That's an important goal, Sykora and Anderson both agree. But they're skeptical it will work. "By and large, with stuff like this, there is no right to counsel," Anderson points out. Besides, Sykora says, many criminal defendants don't even have an accurate understanding of their own criminal records. "We can establish all kinds of ways for these people to challenge the information. But there are enormous numbers of public defense clients who are so low-functioning-- people with fetal alcohol syndrome and other cognitive difficulties--that they won't have any idea whether the record they're looking at is theirs or not."
At the opposite end of the spectrum, says Sykora, there are high-level sociopaths who know how to work the system. "Designing a system to address the Donald Bloms of the world loses sight of the fact that 99.9 percent of the people in the system aren't anywhere near Donald Blom's level of craftiness," Sykora says. "For every Donald Blom, there are thousands of people who can scarcely manipulate their way through the line at the grocery store."
Still, it was the specter of Donald Blom that got the CriMNet train out of the station. The notorious case of 9/11 hijacker Mohammed Atta helped keep it rolling down the tracks. Two months before the World Trade Center attack, Atta was pulled over by police in Palm Beach County, Florida, on a routine traffic stop. At the time, there was a bench warrant for Atta in the neighboring Broward County for failure to appear in court. But because the cops in Palm Beach County didn't know about the Broward County warrant, Atta was released. And we all know what happened next.
Last summer, two of CriMNet's most prominent boosters--Charlie Weaver, then public safety commissioner and now Pawlenty's chief of staff, and Target CEO Robert Ulrich--invoked the Atta incident as part of a CriMNet public relations blitz. In a joint op-ed piece entitled "Mr. Magoo vs. the terrorists" (which initially ran in the Washington Times, then in the Star Tribune), Weaver and Ulrich concluded with a call to arms: "We won't approach the type of safety this nation deserves until tracking criminals and terrorists is as effective as tracking inventory at a Target Store. We need to know exactly where they are."
Interestingly, the piece didn't address the issue of whether or not it would have mattered if the cops had known about the outstanding warrant for Atta. Atta was facing the relatively minor charge of failure to appear on a motor vehicle citation. It seems unlikely he would have been locked up for two months. But that wasn't the real point. What mattered was linking gaps in the criminal justice information system with terrorism, and marketing CriMNet as a possible solution.
The Weaver-Ulrich op-ed pieces were not the first effort to wed CriMNet to homeland security. Shortly after the 9/11 attacks, CriMNet deputy director Kooy authored a seven-page white paper presenting CriMNet as a template to the federal government. Among other things, the paper examined a scenario in which commercial databases with information on airplane ticket purchases and car rentals could be matched against data on CriMNet to root out terrorists. That would work best, Kooy posited, if other states were also to adopt the CriMNet model. "Within this architecture, reporting, collection, analysis, and communication can all be securely integrated, in real time, nationwide."
As public safety commissioner Stanek sees it, linking CriMNet to federal anti-terrorism efforts makes perfect sense. "Homeland security is all about intelligence and information sharing. Well, that's what CriMNet is. It's just a name for intelligence and information sharing," Stanek says. Since the launching of CriMNet development efforts, Stanek says, he's taken about a dozen trips to Washington, D.C. Twice, he's met with Tom Ridge, the secretary of the new Department of Homeland Security. Stanek also has courted officials with the Department of Justice, largely in the hopes of garnering more federal funds for the project. To date, the project has received two federal grants totaling some $5 million. "There's a lot of federal money out there," Stanek adds. "Our job is to bring it back to Minnesota."
Despite all these efforts to market the CriMNet plan on a national scale, it remains an open question whether the system will work. Already, red flags have popped up. Last month the CriMNet Policy Group quietly but unanimously voted not to renew the contract for David Billeter, an attorney and e-commerce specialist from General Motors who was hired to run the project a little more than a year ago. Few people associated with CriMNet would go on the record to discuss the reasons for Billeter's departure, and Billeter could not be reached. In Stanek's view, Billeter was smart and talented, but there was concern about his "political and managerial skills."
Not long after Billeter was hired, state representative Phyllis Kahn sent a memo to fellow legislator Phil Krinkie calling into question some of Billeter's decisions. In particular, Kahn raised her eyebrow at the selection of a small Atlanta-based software company called Mobiam Solutions, Inc. to design the CriMNet backbone. Because Mobiam's CEO, Steve Kostyshen, served as the CEO of another dot-com that had business dealings with General Motors, Kahn figured Billeter should have recused himself from the selection process.
In December, meanwhile, a group of local criminal justice IT professionals met in St. Paul to evaluate the CrimMNet design. After two days, the group forwarded its findings to the CriMNet steering committee. In a draft report, the committee concluded that the CriMNet approach, "while conceptually viable, is very high risk." Echoing Kahn's concerns, the report questioned the selection of Mobiam to design the backbone, noting that the company is "a very small and very new vendor." (The company, founded in 2001, lists only two customers on its website: the State of Minnesota and ACTRA, a Canadian performing-artists union. Neither Mobiam CEO Kostyshen nor the company's public relations representative returned City Pages' calls.) The steering committee report raised other questions, too. Uncertainty about the size and scope of the project, as well as future funding, posed a "catastrophic" risk. In the end, the report recommended that the current design be scrapped and the whole CriMNet system be "reconstituted."
That recommendation was rejected by the CriMNet Policy Group. CriMNet's Kooy points out that the assessment from the steering committee came in a draft report. And many of the conclusions, he says, were based on incomplete information. "There was a lot of concern on our part when we saw those contentions, and quite a lot of hullabaloo," he says. "But I think most of that was ironed out over the holidays."
Others remain doubtful. State Representative Krinkie notes that large-scale government technology projects are notorious for their high failure rates. By some estimates, as many as 85 percent of such projects run aground eventually. In large part, such failures are a product of the design-by-committee approach. Krinkie believes the ambitious nature of the CriMNet model heightens the risk. "Our track record at the state level has been poor, and here we're trying to do something no other state has even bothered to undertake," Krinkie says. "I think we bit off more than we can chew."
Krinkie, who has attained much of his notoriety in the legislature as a keen and vociferous critic of wasteful government projects, contends that CriMNet should be significantly scaled back or possibly scrapped altogether; he believes the prospect of continuing funding with no clear limit in view just doesn't make sense. In the past decade, according to a new report prepared by the House of Representatives Fiscal Analysis Department, the state has spent more than $112 million on criminal justice information projects. The department projects another $86 million in expenditures in the next two years, and, as Krinkie emphasizes, there is no end in sight. "This is a system that could easily run half a billion dollars, maybe even a billion," Krinkie theorizes.
Not surprisingly, the project's supporters roundly dismiss such naysaying. On the issue of cost, they often say, the question is not whether the state can afford to build CriMNet, but whether it can afford not to. And they dismiss concerns about possible design flaws. "Failure is not an option on this project; it's just too important. This is the number one criminal justice project in the state and it's got support across the board. I think within two years you'll see it statewide on a fairly regular basis," says Rich Stanek.
As the new public safety commissioner, Stanek will be a major player in the future of CriMNet. And in his view, the future is bright. Within a few years, he suspects that the general public will have access to one of CriMNet's most valuable features, the criminal history check. And that, he maintains, is only right: "Should I be able to go on the Internet and look up criminal information on the neighbor down the street who wants to baby-sit my kids? Yes."
He will have some interesting bedfellows in making that case. Mark Anfinson, an attorney who represents the Minnesota Newspaper Association (and City Pages), says his organization is watching the discussions about the future of CriMNet with keen interest. And he says that the media will fight for access to CriMNet data. "We fundamentally oppose any exemption, temporary or permanent, to any data that is currently public," Anfinson says.
Won't that expose people to harm, particularly if there is inaccurate information on the system? Anfinson acknowledges the risk.
But, he says, loss of individual privacy may simply be an inevitable outcome of the digital age. In the end, CriMNet will simply be one more manifestation of that phenomenon. Technical glitches or financial constraints may well derail the project in the short term. But that will likely be a temporary setback. In the end, criminal histories--as well as vast repositories of other police intelligence--will be lighting up PC screens across the state.
As Anfinson sees it, the best that can be hoped for is that the government ensures that the information on CriMNet is fair and accurate. And if it remains public, he contends, the government will be more apt to take steps to do just that. Still, Anfinson says, there will be negative consequences. People's past transgressions will follow them, old failings will not be forgotten. "There's no question that the potential for one's past to haunt one's future will be amplified dramatically by this stuff.
"We're all going to be living naked in glass houses," Anfinson says, "and we're just going to have to get accustomed to the view."