Skip to comments.Precursor of Database Project Gave Feds 120,000 Suspects
Posted on 05/19/2004 10:30:38 PM PDT by TheOtherOne
AP Exclusive: Precursor of Database Project Gave Feds 120,000 Suspects
Published: May 20, 2004
The "high terrorism factor" scoring system also became a key selling point for the involvement of the database company, Seisint Inc., in the Matrix project.
Public records obtained by The Associated Press from several states show that Justice Department officials cited the scoring technology in appointing Seisint sole contractor on the federally funded, $12 million project.
Seisint and the law enforcement officials who oversee Matrix insist that the terrorism scoring system ultimately was kept out of the project, largely because of privacy concerns.
However, new details about Seisint's development of the "terrorism quotient," including the revelation that authorities apparently acted on the list of 120,000, are renewing privacy activists' suspicions about Matrix's potential power.
"Assuming they have in fact abandoned the terrorist quotient, there's nothing that stops them from bringing it back," said Barry Steinhardt, director of the technology and liberty program at the American Civil Liberties Union, which learned about the list of 120,000 through its own records request in Utah.
Matrix - short for Multistate Anti-Terrorism Information Exchange - combines state records and data culled by Seisint to give investigators fast access to information on crime and terrorism suspects. It was launched in 2002.
Because the system includes information on innocent people as well as known criminals, Matrix has drawn objections from liberal and conservative privacy groups. Utah and at least eight other states have pulled out, leaving Florida, Connecticut, Ohio, Michigan and Pennsylvania.
The AP has received thousands of pages of Matrix documents in records requests this year, including meeting minutes and presentation materials that discuss the project in detail.
Not one indicates that Matrix planners decided against using the statistical method of determining an individual's propensity for terrorism.
When the AP specifically requested documents indicating the scoring system was scrapped, the general counsel's office for Florida state police said it could not uncover any.
Even so, people involved with Matrix pledge that the statistical method was removed from the final product.
"I'll put my 26 years of law enforcement experience on the line. It is not in there," said Mark Zadra, chief investigator for the Florida Department of Law Enforcement.
He said Matrix, which has 4 billion records, merely speeds access to material that police have always been able to get from disparate sources, and does not automatically or proactively finger suspects.
Bill Shrewsbury, a Seisint executive and former federal drug agent, said the terrorism scoring algorithm that produced the list of 120,000 names was "put on the shelf" after it was demonstrated immediately following Sept. 11, 2001.
He said the scoring system requires intelligence data that was fed into the software for the initial demonstration but is not commonly available. "Nor are we interested in pursuing that," he said.
The Utah documents included a Seisint presentation saying the scoring system was developed by the company and law enforcement officials by reverse engineering an unnamed "Terrorist Handbook" that reveals how terrorists "penetrate and live in our society."
The scoring incorporated such factors as age, gender, ethnicity, credit history, "investigational data," information about pilot and driver licenses, and connections to "dirty" addresses known to have been used by other suspects.
According to Seisint's presentation, dated January 2003 and marked confidential, the 120,000 names with the highest scores were given to the Immigration and Naturalization Service, FBI, Secret Service and Florida state police. (Later, those agencies would help craft the software that queries Matrix.)
Of the people with the 80 highest scores, five were among the Sept. 11 hijackers, Seisint's presentation said. Forty-five were identified as being or possibly being under existing investigations, while 30 others "were unknown to FBI."
"Investigations were triggered and arrests were made by INS and other agencies," the presentation added. Two bullet points stated: "Several arrests within one week" and "Scores of other arrests." It does not provide details of when and where the investigations and arrests occurred.
Phil Ramer, who heads Florida state police's intelligence division, said his agency found the list a useful starting point for some investigations, though he said he could not recall how many. He stressed that the list was not used as the sole evidence to make arrests.
"What we did with the list is we went back and found out how they got on the list," Ramer said.
Dean Boyd, a spokesman for U.S. Immigration and Customs Enforcement, a descendant of INS in the Department of Homeland Security, said he could not confirm that INS used or was given the list.
Although Seisint says it shelved the scoring system - known as high terrorist factor, or HTF - after the original demonstrations in the wake of Sept. 11, 2001, the algorithm was touted well into 2003.
A records request by the AP in Florida turned up "briefing points," dated January 2003, for a presentation on Matrix to Vice President Dick Cheney and other top federal officials delivered jointly by Seisint, Florida Gov. Jeb Bush and Florida's top police official.
One of the items on Seisint's agenda: "Demonstrate HTF with mapping." Matrix meeting minutes from February 2003 say Cheney was briefed along with Homeland Security Director Tom Ridge and FBI Director Robert Mueller.
In May 2003, the Justice Department approved Seisint as sole data contractor on the project, citing the company's "technical qualifications," including software "applying the 'terrorism quotient' in all cases."
"The quotient identifies a set of criteria which accurately singled out characteristics related to the perpetrators of the 9-11 attacks and other terrorist events," said a memo from an Office of Justice Programs policy adviser, Bruce Edwards. "This process produced a scoring mechanism (that), when applied to the general criminal population, yields other people that may have similar motives."
A spokeswoman for the Office of Justice Programs declined to comment.
Ramer, the Florida agent, said the scoring system was scrapped because it was "really specific to 9/11," and not applicable for everyday use. Also, he said, "we didn't want anybody abusing it."
Seisint Inc., is a Boca Raton, Fla., company founded by a millionaire, Hank Asher, who stepped down from its board of directors last year after revelations of past ties to drug smugglers.
AP Investigative Researcher Randy Herschaft contributed to this report.
On the Net:
When we take the next big hit people will be outraged that this system was NOT used.
| The 20 hijackers are known...and they only get a 25% hit rate ? Is that good ?
We can't tell from anything written here. Seisint's database is basically public records data plus credit bureau stuff. I would bet that the Amazing Terrorism Quotient was thrown together by Hank Asher himself in less than a day, probably just for the hell of it. It probably had no more "statistics" in it than some weighting scheme that started out as WAGs. The weights were then diddled with until all the hijackers in the data base floated to the top. Presto, a "product" for the feds.
What they don't tell us is what happened when they looked into the other 95 people in the top 100. Have they uncovered any other plots with this, or just a bunch of recent arrivals with Arabic-sounding names who happened to live in Arab neighborhoods? My guess would be the latter.
There is probably some merit in a Bayesian network type deal involving the bad guys' addresses and acquaintences, but I doubt that was in anything Seisint did (at least that early on -- they might be doing it now).
The biggest problem with all these statistical methods for finding terrorists by data-mining U.S. public records and credit data is that the bad guys have very few footprints. It's a wonder they even found 5 of them in there... it was probably the credit cards that got them in.
If you are a recent arrival to America, and you don't purchase any real proerty, and you don't get yourself arrested, and you stay away from consumer credit, you could stay off Seisint's (or ChoicePoint's, or any of those guys') radar for a long time... certainly long enough to plan and execute a Bad Thing.
Plus, the terrorist algorithm could be adapted to fight crime in other ways too! Like predicting if someone is a likely speeder, or if they thing "strange" thoughts, or if they're an enemy of the administration.
Anyhoo, once computer code is created it never goes away. It's not like breaking a mold.
You are wise beyond your beers. :)
A good example, Governors can sign laws that say the police/government can't keep databses of gun-owners, like Jeb Bush in this thread on FR, but that doesn't mean it won't happen in another form (such as magazine subscription lists, credit card statements, club memberships, etc.). After all, if guns are used to commit crimes, shouldn't the government know where they are all at.
Funny you mention about the databases predicting if somebody is an enemy of the administration. If we get any more PATRIOT laws passed, pretty soon the government can officially monitor various groups on the internet to see what their intentions are. I'm sure the liberals would consider FReepers/Conservatives to be enemies of an administration (and armed ones at that).
A gut feeling said that when they killed TIA, etc., that they weren't really killing it. It's like a Hydra, cut one head off, two more pop up.
The sad thing is, if there was another terrorist attack on US soil (although actually the usage of the word "terrorist" is being expanded in scope), all the people who rely on the government to take care of them would really start making a lot more noise, and would have no problem giving up what little privacy we have left, and would probably give up quite a bit of freedom, just to have the government's blanket of protection wrapped around them (How's that for a run-on sentence, lol). A typtical liberal reaction is to run to the government, and with apparently half the country leaning that way, things could get ugly, especially for Conservatives.