Abstract

Excerpted From: Ande Davis, A Preponderance of Bias: Why Artificial Intelligence Should Be Qualified Immunity's Fatal Flaw, 61 Washburn Law Journal 565 (Spring, 2022) (299 Footnotes) (Full Document)

 

AndeDavisIn the wake of the 2020 police killings of Breonna Taylor in Louisville, Kentucky, and George Floyd in Minneapolis, Minnesota, the public discussion of criminal accountability for law enforcement was accompanied by a related discussion around civil remedies for victims. This secondary discussion brought new public attention to the impediments posed by the doctrine of qualified immunity. News media devoted significant amounts of coverage to the protests throughout the following months, as well as to the corporate responses and government actions that followed; however, both the conversation around disparate treatment of Black people and minorities by law enforcement and the criticism of qualified immunity existed long before the 2020 national unrest brought them into the public spotlight.

Since the mid-2010s, the Black Lives Matter movement has platformed the discussion around police violence that first gained national prominence during the responses to the 2014 killings of Michael Brown in Ferguson, Missouri, and Eric Garner in New York. Less prevalent in the media, a number of legal battles over the past few decades have continued to strengthen and shape the qualified immunity doctrine despite two of the doctrine's most influential critics in the Supreme Court--Justices Clarence Thomas and Sonia Sotomayor-- authoring multiple recent rebukes of the doctrine in their opinions. It is unsurprising that the twin issues of police overstepping their duty to public safety and the doctrine of qualified immunity would coexist in the public discourse amidst a national outcry over disparate racial treatment by law enforcement. Both share roots in the history of "law and order" policymaking that has emboldened and endorsed aggressive law enforcement actions for much of U.S. history, and certainly since the mid-twentieth century.

The protests that started in the summer of 2020 have also encountered a new problem in policing that policymakers of decades past had not foreseen: the introduction of artificial intelligence ("AI")--machine learning, predictive algorithms, and facial recognition, among other systems--into law enforcement practices. These algorithms, particularly those that have been adopted by law enforcement agencies and court systems, have demonstrated an engrained bias against people of color, and particularly toward Black people. For instance, some AI systems with law enforcement and criminal justice applications include risk assessment software to assist in determining jail sentences based on expected recidivism. This expected recidivism is determined from preexisting datasets in which Black individuals were overrepresented or facial recognition software that frequently misidentifies Black people due to an underrepresentation in training datasets. While a number of technology companies have released statements postponing their facial recognition software development, abandoning it altogether, or barring law enforcement from using their software, many agencies and courts have already implemented some form of facial recognition or other AI into their investigation, sentencing, and parole practices. These adoptions create the potential for further violations of due process and Fourth Amendment rights, particularly the rights of those negatively affected by systemic bias.

At the same time machine learning algorithms provide additional opportunities for law enforcement to violate individuals' constitutional rights, the doctrine of qualified immunity imbues AI as an avenue through which law enforcement can avoid civil accountability. Many of these technologies are simply too new to provide a clearly defined constitutional violation required by the qualified immunity doctrine. Further, a broader difficulty remains in protecting individuals' constitutional rights from AI systems due to those systems' lack of provable intentionality and the dispersed responsibility between the systems' design, adoption, and use. The intersection of these two issues stands to have the greatest impact on minority--and particularly Black-- communities because of the disproportionate potential for them to be victims of bias from both AI and law enforcement, while simultaneously limiting avenues for civil recourse. Numerous arguments have previously been advanced for the abolition of qualified immunity as a doctrine, but the growing implementation of AI provides distinct illustrations for why qualified immunity's usefulness in law is outdated and unworkable in practice.

Section II will discuss the racialized roots of qualified immunity in "law and order" politics meant to imbue police with more resources and authority as a countermeasure against the growing Black political and social agency in the midst of the Civil Rights Movement and beyond. Section III will explore the implementation of AI mechanisms and the prevalence of racial bias in the growing set of AI tools employed by law enforcement. Section IV will then analyze the impact that AI can have on the qualified immunity doctrine and how the doctrine's development presents a particular obstacle to Black communities and individuals in gaining civil redress for violations of their constitutional rights. That section argues that AI's implementation casts qualified immunity's existing problems into stark relief and strengthens the list of arguments for why the doctrine should be abolished.

[. . .]

Critics of qualified immunity have been numerous. In a recent series of articles published across several journals, Joanna C. Schwartz has pointed out numerous grounds upon which arguments in favor of qualified immunity fail. First, she has pointed to its legally dubious foundations in highlighting that the doctrine has no basis in common law, despite the Court's assertion that it did in Pierson v. Ray. Second, she has made multiple arguments that qualified immunity simply fails on its face by not accomplishing what the numerous Justices over the years have offered as reasons for its existence--she has demonstrated that it does not suffice to shield officers from financial burdens or the burdens of litigation, that it protects neither government budgets nor constitutional rights, and that it frequently does not function as the reason to dismiss a sufficient number of lawsuits against law enforcement officers. Finally, she has made the case that qualified immunity undercuts constitutional protections.

Karen M. Blum highlights similar problems with qualified immunity, both practical and constitutional, namely that it prevents articulation of constitutional law by allowing courts to skip to the second prong of the test, permits interlocutory appeals that drain resources of both the litigants and the courts, and muddies up legal principles by providing too much discretion to the judge and thereby depriving plaintiffs of the right to a trial by jury. Meanwhile, Aaron Nielson and Christopher Walker make the argument that, while qualified immunity is a creature of the Court, its adoption has created entanglements with federalism such that any proper addressing of the doctrine is best handled as a legislative matter rather than a judicial one.

Several lobbying organizations have gotten involved in the fight against qualified immunity. The libertarian Institute for Justice has made similar arguments, stating that "if a government official finds a new and unique way to violate someone's constitutional rights, there is little that can be done to hold the official accountable," at least in civil court. Another libertarian organization, the Cato Institute, "has launched a full blown assault on the doctrine of qualified immunity," hosting webinar panels of scholars and legal professionals to attack the foundations of the doctrine and filing amici briefs on cases involving qualified immunity. Cato has levied numerous attacks against the doctrine, writing about "the absurdity and injustice of the 'clearly established law' standard that characterizes modern qualified immunity doctrine," observing that "qualified immunity was invented by the Supreme Court out of whole cloth and has no basis [in] statutory text, legislative intent, or sound public policy," and pointing out that "a troubling proportion of the statements that the law enforcement lobby makes in support of qualified immunity are not just misguided or misleading, but outright false." Meanwhile, the liberal American Civil Liberties Union ("ACLU") has similarly weighed in on the issue:

We have seen the deadly consequences [of police violence] play out on the streets, and Black Americans have largely paid the price. Recent events demonstrate the urgent need for Congress to stand up for the rule of law and abolish qualified immunity--for anyone acting under color of law--to close the loophole allowing government officials to escape accountability for violating constitutional rights.

Like the Cato Institute, the ACLU has also filed amici briefs on behalf of litigants in suits involving qualified immunity defenses, occasionally in tandem with Cato and other groups.

Legislation has been introduced in both the House and Senate to make legislative alterations to qualified immunity in response to criticism of the doctrine. In 2020, Libertarian Representative Justin Amash and Democratic Representative Ayanna Pressley introduced the Ending Qualified Immunity Act in the House of Representatives, which would remove the ability for a defendant acting under the color of law to assert a defense that they were acting in good faith or believed that their conduct was lawful at the time it was committed. Republican Senator Mike Braun introduced competing legislation in the Senate--the Reforming Qualified Immunity Act--under which any law enforcement officer who wanted to assert the qualified immunity defense would bear the burden of proving that their conduct was authorized by a statute or court case in the relevant jurisdiction, effectively flipping the presumption established by the Supreme Court. Any of these acts would substantially alter both the effect and understanding of § 1983 and could change the landscape of civil rights litigation. What effect any legislation of this type might produce on the actions of law enforcement would likely be part of any debate to be had on the legislative floor, if the bill were to make it that far.

The Court has not indicated any preference for scaling back or repealing the doctrine. Biden signaled during the 2020 campaign that he was open to reforming the qualified immunity doctrine, although in the first several months of his administration he had not made any serious attempts at doing so. Many of the legislators who introduced bills repealing or altering qualified immunity remain in office, so the potential remains for debate over the doctrine to rage on in the halls of Congress. Absent one of these bodies taking action on qualified immunity, the task may be left to state legislators to address the issue, which could result in patchwork results across the country regarding the success of § 1983 claims. The growth of AI technologies within law enforcement has illuminated the writing that has long been on the wall. Because many prosecutors refuse to hold law enforcement criminally accountable for constitutional violations, qualified immunity must be rescinded or the civil protections afforded to individuals in the Constitution are in imminent danger of being severely rolled back.

The 1871 Civil Rights Act was implemented as a safeguard to Black communities whose rights, lives, and liberty were being threatened by members of the Ku Klux Klan acting under the color of law, often with either the tacit permission or active participation of local law enforcement. Because of the reluctance of White society to fully protect the constitutional rights of Black citizens, the Act was rarely enforced, leading Black citizens to fight for their own civil rights during the first half of the twentieth century. In the wake of the social upheaval that came along with Black people claiming their due constitutional rights, the same reluctant members of White society incited calls for "law and order" to tamp down the progress being made by Black communities. In that "law and order" spirit, the Court offered protection to law enforcement in the form of qualified immunity, shielding them from civil liability through a doctrine that continued to expand its reach over the next several decades. More recently, police have obtained new machine learning tools that have been demonstrated to frequently show bias toward Black people due to training data rooted in historical racism and poor performance in recognizing dark-skinned faces.

The coalescence of all these issues casts a pall over the nation's Black population. Law enforcement using AI shown to be biased against Black communities can then be shielded from Black plaintiffs seeking recourse for violation of their constitutionally granted civil rights by invoking qualified immunity. This situation is evidence that technology has outpaced any perceived usefulness of qualified immunity as an effective tool. The simultaneous existence of both qualified immunity and AI law enforcement tools may very well prove untenable when courts are given wide discretion on the need to rule whether a law enforcement action indeed violates a constitutional right. Given society's interest in enhancing the efficiency of law enforcement and the economic interests of companies building AI technologies, it is unlikely that AI will disappear from law enforcement's toolbox. In order to ensure that individuals--and Black people, in particular--are not subject to an erasure of their constitutional rights at the hands of law enforcement personnel, the doctrine of qualified immunity needs to be overturned. Its ability to be applied too broadly has been bolstered by the application of tools too new to pass the test necessary for a claim to move forward, effectively undoing any of the necessary social good provided by § 1983.


B.A. English 2006, Washburn University; M.F.A. Creative Writing 2009, Minnesota State University, Mankato; Ph.D. Candidate, English and Humanities 2021, University of Missouri-Kansas City; J.D. Candidate 2022, Washburn University School of Law.