Unreliable ShotSpotter Audio Previously Used to Convict Chicago Man in Murder Case

ShotSpotter's acoustic gunshot detection system was used by Chicago police as critical evidence in a murder case, leading to the wrongful arrest and year-long incarceration of a 65-year-old man. Investigations revealed that the system frequently misclassifies sounds and that company employees manually modify alerts at the request of police, introducing human bias. The case highlights concerns regarding the lack of transparency, reliability, and potential for racial bias in algorithmic policing tools.

ShotSpotter audios were previously admitted to convict an innocent Black man in a murder case in Chicago, resulted in his nearly-one-year-long arrest before being dismissed by prosecutors as insufficient evidence.

Source: AI Incident Database

Risk classification

  • Primary risk domain: 7 AI system safety, failures, & limitations
  • Primary risk subdomain: 7.3 Lack of capability or robustness

The core failure stems from the AI system's inability to reliably distinguish gunshots from other percussive sounds like fireworks or engine backfires, leading to highly inaccurate alerts.

Additional risk subdomains

  • 1.1 Unfair discrimination and misrepresentation: The system's sensors are deployed almost exclusively in Black and Latino neighborhoods, leading to systemic over-policing and discriminatory enforcement.
  • 5.1 Overreliance and unsafe use: Police and prosecutors placed blind faith in the technology as direct evidence, ignoring disclaimers that it should only be used for initial investigative leads.

Causal factors

  • Entity: AI
  • Intent: Unintentional
  • Timing: Post-deployment

The incident was primarily caused by the lack of capability and high false-positive rate of the deployed ShotSpotter AI system, which misclassified ambient noises as gunshots.

EU AI Act risk tier

  • Risk tier: 2 High Risk

High Risk: The system is used in law enforcement applications, specifically for crime detection and dispatching, which has significant implications for public safety and fundamental civil rights.

AI system and alleged parties

  • AI system: ShotSpotter
  • AI purpose: Gunshot Detection; Threat Detection
  • Behaviour type: Assistant
  • Alleged developer: ShotSpotter
  • Alleged deployer: Chicago Police Department
  • Alleged harmed parties: Michael Williams

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

  • Physical: direct Negligible, indirect Substantial
  • Infrastructure: direct Negligible, indirect Negligible
  • Property: direct Negligible, indirect Negligible
  • Financial: direct Negligible, indirect Minor
  • Environmental: direct Negligible, indirect Negligible
  • Malicious content: direct Negligible, indirect Negligible
  • Differential treatment: direct Negligible, indirect Substantial
  • Civil rights: direct Negligible, indirect Minor
  • Democracy: direct Negligible, indirect Negligible
  • Privacy: direct Negligible, indirect Substantial
  • Psychological: direct Negligible, indirect Substantial
  • Epistemic: direct Minor, indirect Substantial
  • Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Physical

Reported: The report explicitly describes physical harm and a fatality indirectly linked to the system's dispatches.

Indirectly caused: Michael Williams developed an uncontrollable physical tremor in his hand during his 11-month incarceration. Additionally, a ShotSpotter alert dispatched police to a scene where they fatally shot 13-year-old Adam Toledo.

Inferred additional harm: Given the high volume of false alerts leading to hostile police deployments, it is highly likely that other individuals suffered physical injuries during wrongful stops or detentions.

Financial

Reported: The report explicitly mentions financial losses incurred by the victims of wrongful arrests.

Indirectly caused: Michael Williams suffered a loss of income and incurred significant legal bills due to his 11-month incarceration.

Inferred additional harm: Many of the other individuals stopped or arrested due to the thousands of dead-end deployments likely suffered financial losses from legal fees and missed work.

Differential treatment

Reported: The report explicitly describes systemic differential treatment of racial minorities.

Indirectly caused: ShotSpotter sensors were deployed almost exclusively in predominantly Black and Latino neighborhoods, exposing these communities to disproportionate police surveillance and hostile encounters.

Inferred additional harm: Systemic racial bias in law enforcement actions is reinforced by the biased geographic placement of the technology.

Civil rights

Reported: The report explicitly describes widespread violations of civil and constitutional rights.

Indirectly caused: Michael Williams and Daniel Ortiz suffered violations of their Fourth Amendment rights through illegal stops, searches, and arrests without probable cause.

Inferred additional harm: Thousands of Chicago residents stopped during the 40,000+ dead-end deployments likely had their civil liberties and constitutional rights violated.

Privacy

Reported: The report explicitly describes privacy violations through surveillance and illegal searches.

Indirectly caused: The continuous acoustic monitoring of neighborhoods without resident consent, and the illegal search of Daniel Ortiz's vehicle by responding officers.

Inferred additional harm: Continuous acoustic surveillance of millions of residents in covered areas without public oversight or consent.

Psychological

Reported: The report explicitly describes severe psychological distress and mental health struggles resulting from wrongful incarceration.

Indirectly caused: Michael Williams suffered severe mental anguish, trauma, and suicidal ideation, planning to take his life with stockpiled pills during his 11-month wrongful imprisonment.

Inferred additional harm: Other individuals subjected to arbitrary arrests and aggressive detentions, such as Daniel Ortiz, likely experienced significant psychological trauma and distress.

Epistemic

Reported: The report explicitly describes the creation of false or altered forensic evidence.

Directly caused: ShotSpotter generated altered forensic reports that misrepresented the nature and location of sounds.

Indirectly caused: Prosecutors introduced these altered reports as scientific evidence in court trials, undermining the integrity of the judicial process.

Inferred additional harm: Numerous other criminal cases may have been compromised by altered or fabricated ShotSpotter evidence that went unchallenged.

People affected

  • Occurrences reported: 4
  • People reportedly harmed: 5
  • People reportedly exposed: 100000

Potential causes

Management

  • Prioritizing Sales Over Accuracy: Accuracy guarantees were made by sales rather than engineers.
  • Secret Contract Extensions: City renewed contracts without public input or oversight.
  • Biased System Deployment: Sensors placed predominantly in minority neighborhoods.

Technology

  • Flawed Classification Algorithm: Algorithms misclassify fireworks or car backfires as gunshots.
  • Proprietary Black Box Software: No independent testing or peer review of the algorithm is allowed.

Data Inputs

  • Acoustic Noise Interference: Environmental sounds like fireworks and helicopters mimic gunshots.
  • Incorrect Initial Geolocation: System generated addresses up to a mile away from actual coordinates.

Human Factors

  • Analyst Manual Overrides: Analysts manually reclassified firecrackers to gunshots.
  • Police Tunnel Vision: Officers put blind faith in alerts and ignored other leads.
  • Lack of Formal Expert Training: No official training materials existed for forensic experts.

Process and Methods

  • Post-Hoc Report Alterations: Analysts modified alerts and shot counts at police request.
  • Misuse of Investigative Leads: Using alerts as primary trial evidence despite disclaimers.
  • Contaminated Feedback Loops: Training models with subjective or incorrect police feedback.

Regulatory Environment

  • Lack of Independent Auditing: No regulatory requirement to audit forensic AI technology.
  • Absence of Standards for AI Evidence: Courts admit unvetted algorithmic evidence without validation.

Information quality

  • Classification confidence: High
  • Reason for confidence: The reports provide extensive, high-quality information including court documents, expert witness testimonies, nonpartisan watchdog investigations, and detailed case studies of affected individuals.
  • Ambiguities identified: The exact proprietary algorithms and source code of ShotSpotter remain a 'black box' and are not publicly auditable.
  • Alternative interpretations: ShotSpotter maintains that its system is 97% accurate and that human analyst overrides are corrections rather than improper alterations.

The failure of ShotSpotter's acoustic gunshot detection system in Chicago led to a wrongful arrest and highlighted systemic issues with algorithmic policing tools. While causing localized civil liberties and human rights concerns, the incident has minor implications for national security.

  • Overall national security impact: Minor
  • Scope: Single nation
  • Primary target: United States
  • Alleged perpetrator: Unknown

Threat characteristics

  • Imminence: Long-term. Represents an ongoing systemic concern regarding algorithmic policing tools rather than an immediate national security crisis.
  • Autonomy: Human-supervised. The AI system detects and classifies sounds autonomously, but requires human analysts and law enforcement officers to verify and act on the alerts.
  • Novelty: Established threat. Algorithmic bias, false positives in automated detection systems, and forensic tool failures are well-documented and established issues.

Impact by dimension

  • Physical security: Negligible. The incident involves a domestic acoustic monitoring system for law enforcement and does not impact critical national infrastructure, kinetic military systems, or physical security at a national level.
  • Information security: Negligible. No classified intelligence systems, espionage, or state-sponsored information warfare campaigns were involved in this municipal policing tool failure.
  • Sovereignty: Minor. Local judicial and law enforcement decision-making processes were compromised, leading to wrongful incarceration, but this does not threaten national sovereignty or core federal government functions.
  • Economic security: Negligible. The failure of a commercial acoustic detection system does not threaten strategic national industries, economic stability, or critical technological advantages.
  • Societal stability: Minor. The system's deployment led to localized civil rights violations, systemic over-policing in minority neighborhoods, and wrongful detentions, representing a minor domestic societal impact.
Explore in the interactive Incident Tracker