SMU Law Review
Abstract
The policing of dissent in America is undergoing a radical transformation that threatens the fundamental right to protest. Demonstrations are no longer merely surveilled; they are algorithmically suppressed. Law enforcement agencies increasingly deploy algorithmic social media surveillance to predict protest violence, forecasting potential “riots” and justifying militarized responses before a single demonstrator takes to the street. This Article identifies and challenges what it terms “predictive protest policing,” arguing that these systems do not objectively detect threats but actively construct them, disproportionately targeting racial justice demonstrations. Drawing on evidence from the federal response to immigration protests in Minneapolis, the preemptive policing of Gaza War protests, and the suppression of Black Lives Matter demonstrations, the Article shows how algorithmic surveillance embeds racial biases into a self-reinforcing feedback loop: protests by racialized communities are flagged as dangerous, met with escalated force, and the resulting confrontations are cited as validation of the original prediction—producing the very violence the technology claims merely to forecast. Although researchers have revealed how surveillance technologies perpetuate racial inequality, legal scholarship has yet to address how predictive AI specifically undermines racial justice advocacy. This Article fills that gap with the first sustained constitutional analysis of predictive protest policing, arguing that these systems threaten First Amendment freedoms of speech, assembly, and association; facilitate suspicionless surveillance prohibited by the Fourth Amendment; perpetuate badges and incidents of slavery under the Thirteenth Amendment; and raise equal protection and due process concerns under the Fourteenth Amendment.
Recommended Citation
Christian Powell Sundquist,
Predictive Protest Policing,
79
SMU L. Rev.
623
(2026)
