JUSTICE_BY_ALGORITHM: Rejected by Iowa Judges
There has been a lot of talk about the use of a pretrial risk assessment tool, a criminal justice algorithm, in deciding who gets bail in Iowa and what conditions of release to which a defendant will be subjected.
In fact, there is currently under operation a four-county pilot project, including Polk County, to deploy the Laura and John Arnold Foundation’s Public Safety Assessment tool, a new algorithm to help judges make what some judges say are better decisions. The tool was developed by the Foundation, which was created by former-Enron executive and multi-billionaire and public-policy meddler John Arnold, who wants to “moneyball” criminal justice in Iowa and collect as much data about Iowa citizens as he can to help his national moneyballing efforts. In fact, the four counties conducting the pilot project signed a contract with the Laura and John Arnold Foundation to use the Public Safety Assessment tool, but then also to conceal it from public view, allow wide access to data including private data about Iowans, and also with the understanding that judges would follow the recommendations of tool eighty-percent of the time.
So much for judicial discretion.
During the last legislative session, however, based on examples from Iowa and other states where high-risk defendants were getting out of jail free, including prior felons in possession of firearms, both the Iowa House and Senate passed legislation that would have immediately sent John Arnold and his Public Safety Assessment tool packing. The footnote in the budget would have ended the Arnold tool immediately. Governor Reynolds, however, line-item vetoed that language, and instead decided that John Arnold could continue his moneyballing, but only until December 31, 2018. After that time, there would have to be some substantial proof that the project is working, some sunshine, and some legislative involvement if the use of the Assessment tool was to continue.
Enter Polk County Chief Judge Gamble to prove the pilot project in Polk County, despite widespread legislative and other concerns with it, is actually working. He is apparently a true believer in John Arnold’s moneyballing, but, of course, is also concerned with the realistic argument that Arnold’s moneyballing is taking away from something that is the cornerstone of American justice—the application of impartial judicial discretion to the facts and circumstances of each case. So, Judge Gamble recently penned an op-ed, where he attempted to point out that the pilot project was a raging success. In so doing, however, Judge Gamble actually pointed out that the project is a complete failure.
Let’s examine why.
It is posited that the Arnold Foundation tool is scientific, where one would assume a bunch of data scientists with white lab coats and thick glasses have run regressions around the world at the finest institutions to make sure the thing is dead on accurate. Instead, the Arnold Foundation relies on its own internal hack data scientists at fake institutes like the Center for Justice, Safety, and Resilience to conduct the studies. What is the Center for Justice, Safety, and Resilience? Google it. There is no Center webpage. There is no discussion of who funds the Center. There is of course a director of the Center, who works for a company called RTI, Inc., where apparently the Center is located. There are no professors in residence at the Center. There is no description of the work of the Center, what the Center does to avoid conflicts of interest, whether any of the Center’s publications must be peer-reviewed prior to release, and most importantly whether the Center acts as hired fake academic guns for its donors. Yet, we need not speculate as to the answer. The Arnold Foundation, according to records we received, has paid RTI, Inc. $761,381 since 2015 for "criminal justice reform" to keep the nonexistent lights on. They are hired guns indeed, who issue various reports saying the Arnold tool works and is not discriminatory.
Nonetheless, let’s assume for the moment that the tool is accurate to a reasonable degree of scientific certainty and is as unbiased and effective as John Arnold’s hired guns say it is. Judge Gamble pointed out that during the pilot project in Polk County the judges receiving recommendations from the tool actually believed otherwise. While he claims this is a major success for the protection of judicial discretion because judges only followed the results of the tool 26% of the time, the bigger reality is that judges rejected the recommendations of John’s Arnold’s tool 74% of the time. In other words, Iowa judges did not think the tool was evidence based at all. Further, as Judge Gamble noted, Judges imposed a monetary bail in 36% of all cases in the pilot study, which is another huge rejection of the Arnold Foundation tool which never recommends the setting of bail in any case.
Judges clearly disagree with the fundamental assumptions made by the Arnold Foundation Risk Assessment Tool, not to mention rejecting the recommendations it made 74% of the time.
As the legislature re-convenes and begins asking whether John Arnold should be allowed to continue his moneyballing in Iowa, the answer must be clearly be no. As an old saying goes, the juice simply ain’t worth the squeeze. Certainly, Judges in Polk County didn’t think so—they thought the Arnold Foundation Public Safety Assessment would have gotten it wrong 74% of the time, and thus would have harmed the public or reduced criminal accountability.
So the question is, why should Iowa continue using such a tool when not only does it not work, one-hundred ten national civil rights groups have called for the end of such pretrial risk assessments throughout the United States due to legitimate concerns that tools like John Arnold’s magnify racial bias in the system. The answer is, Iowa should send John Arnold’s tool packing. Yet, knowing John Arnold, we’re sure we’ll soon see a new study proclaiming the success of his Iowa moneyballing—even though we know now that learned judges in Iowa think otherwise.
ResearchCA - Proposition 25 Charitable Bail Funds COVID-19 Does Bail Reform Increase Crime? Bureau of Justice Statistics Costs of Pretrial / Bail Reform Performance of Bail Effectiveness of Bail Preventative Detention Risk Assessments Dr Krahl - Univ of Tampa Brennan Center Arnold Foundation