As the conversation continues on Idaho’s handling of the pretrial release of criminal defendants and the role of algorithm based pretrial risk assessment tools, so do the usual misguided arguments.
Enter former Ada County Sheriff Gary Raney.
Sheriff Raney thinks “black box” risk assessment tools work in an article published this morning (Guest opinion: Using risk assessments in the criminal justice system). Missing in the article's published bio of Sheriff Raney is the fact that he’s on the Board of Directors and the former Chairman of the Board of the Pretrial Justice Institute, a Maryland-based think tank that advocates for the use of computer algorithm driven pretrial risk assessments as an alternative to financially secured bail. Of course, significant funding for the Pretrial Justice Institute comes from billionaire George Soros who wants to take a wrecking-ball to America’s criminal justice system.
House Bill 118, introduced by Representative Greg Chaney, is legislation aimed at reining in computerized pretrial risk assessments by installing measures to prevent racial and gender discrimination during the process of pretrial release. In addition, the proposed bill seeks to force the transparency of these tools and open them up to public inspection.
In the article, Mr. Raney weighs in on House Bill 118, by first erroneously claiming the legislation would stop the use of risk assessments in Idaho. He also failed to point out to the readers that he has a fiduciary duty by virtue of his board position at the Pretrial Justice Institute to advance pretrial risk assessments in Idaho and throughout the United States of America. In fact, tax records from the Pretrial Justice Institute show he has devoted at least 10 hours a week for many years in the pursuit of advancing the use of pretrial risk assessments, no doubt by using his former law enforcement status.
In his op-ed, Raney actually says that House Bill 118 would “ban the use of risk assessments in the criminal justice system.” We’re not sure exactly where he’s getting that, but that is false. First, the legislation only concerns “pretrial risk assessments,” and would not touch algorithms used for policing, probation, parole, sentencing, etc. Second, the legislation doesn’t ban anything. It simply requires that the risk assessments be tested for racial bias and to be shown that they are free of bias, i.e., that they don’t make the system worse. Third, the legislation requires the algorithms to be transparent so that the public, district attorneys, judges, defenders, and criminal defendants may understand how they work and put some sunshine on them in the process. Not exactly a ban, as Mr. Raney suggests.
Now, let’s look at Mr. Raney’s defense of the criticism of the risk assessments: that they could be potentially biased and that they don’t work. Rather than require an algorithm to be found free from bias (under any reasonable statistical approach to that of which there are several), Mr. Raney defends not opening the door to investigate, pursue and stop bias by saying that, “The proposed legislation asserts that assessments like these should not be used because they are racially biased. That view fails to consider the greater bias in the system when there is not an assessment.” In other words, he is arguing that the risk assessment reduces racial bias. What evidence does he have of that conclusion? Apparently, he disagrees with 110 national civil rights groups, not in 2016, but in 2018 who believe they are biased and should be banned across the United States of America.
Of course, Mr. Raney does refer to the landmark, peer-reviewed article by Professor Megan Stevenson of the Antonin Scalia School of Law at George Mason University who found that risk assessments in practice failed to achieve their goals by any measure and have not made the system any better, despite the millions upon millions of dollars spent on such policies.
In addition, if the risk assessment process is less racially biased, what is Mr. Raney afraid of? Wouldn’t he support the legislation if he knew the tools served to reduce bias – which he apparently he believes. Why would he be against it? What does he or the Idaho Supreme Court have to hide? In reality, he knows the tools could actually make it worse, and at a minimum they haven’t been tested for that, but because he has to advocate that they should somehow replace the right to bail, it’s no wonder he takes to spilling ink in the papers in favor of his cause in an alarmist and inaccurate fashion designed to misinform the legislature and local officials.
On the issue of transparency, Mr. Raney says nothing to defend the secrecy of algorithms in Idaho. Indeed, if there is “ample research” supporting that the tool being used in Idaho makes the system less biased as Mr. Raney suggests, then we should all get to see it. Independent researchers should be allowed to come in and test these algorithms, and not leave it to Gary Raney’s pretrial ninjas down at the Pretrial Justice Institute to tell us everything is great and it works. If it’s that “scientific” then removing the cloak of secrecy will immediately get the public the truth.
Interestingly, there is a serious conversation going on regarding risk assessments in other states that Mr. Raney and his team at the Pretrial Justice Institute seem to want to sweep under the rug. In Iowa, pretrial risk tools got the boot. In California, civil rights groups opposed the bail reform legislation, citing the application biased pretrial risk assessment tools. The list goes on…
These tools tend to reinforce the system’s ingrained biases and lack transparency. The data they use, especially arrest and conviction history is greatly skewed by racial and class bias in policing and court outcomes and societal inequities. Even proponents of the tools acknowledge that they will, at best, reflect these biases. Others fear that they will entrench and increase these inequities. – Human Rights Watch
Regardless of whether House Bill 118 becomes law, this is a critical conversation to have and Mr. Raney’s flimsy defense of such a policy is both misguided and more evidence that something does indeed stink.