AI can lead to false arrests and wrongful convictions

Source: Steven L. Johnson University of Virginia

Steven L. Johnson is an Associate Professor of Commerce at the University of Virginia's McIntire School of Commerce, with research published in top-tier journals like MIS Quarterly, Organization Science, and Information Systems Research. He has received prestigious awards such as the Association for Information Systems Distinguished Member Cum Laude (2021) and MISQ Best Paper Award (2020). His expertise in online communities, social network analysis, and digital technology impacts is supported by editorial roles and academic leadership positions.

Source: Salon

Salon.com is an American news and opinion website launched on April 18, 1995, publishing articles on U.S. politics, culture, and current events, with documented critiques including a retracted vaccine conference article and biased coverage of topics like Otto Warmbier and Ron DeSantis. Its homepage features opinion pieces highly critical of Republicans and Donald Trump, such as 'The dark purpose behind Trump's Washington makeover.' Owned by Find.co, it aligns with progressive viewpoints, potentially prioritizing narrative over neutral reporting.

Source: Maria Lungu University of Virginia

Maria Lungu is a 2024-2026 Postdoctoral Research Fellow in the Digital Technology for Democracy Lab at the University of Virginia's Karsh Institute of Democracy, specializing in AI's intersection with law, public policy, criminal justice risk-assessment tools, policing, and predictive policing impacts. Her work has been published in the Journal of AI Law and Regulation. As an academic researcher, her output is confined to scholarly contexts with no evidence of media production or fact-checking ratings.

Source: The Conversation

The Conversation is a network of nonprofit media outlets that publishes news stories and research reports online, authored by academics and edited by professional journalists to ensure academic rigor and journalistic flair. Articles are written by researchers from institutions like Harvard Kennedy School, University of Maryland, and Indiana University, covering topics such as politics, science, and health. No specific fact-checking track record or ratings are mentioned.

Factual Error

Article states "In Baltimore on Oct. 20, 2025, a 17-year-old student named Taki Allen" but incident was in Baltimore County at Kenwood High School, and Allen was 16.

Mislocating to "Baltimore" (city) instead of Baltimore County implies urban crime context, potentially biasing perception of environment.

unverified_claim

"Research at the University of Virginia’s Digital Technology for Democracy Lab with police chiefs shows that some law enforcement groups follow strict policies... while others have no such policy."

Presents specific research as evidence without confirmation, lending undue authority to claims about varying police AI policies.

unverified_claim

"Some researchers have argued that predictive policing systems do not increase the likelihood that racial minorities will be arrested more often relative to traditional policing practices."

Uses this to pivot from disparities concern, but no supporting researchers found; most sources highlight bias risks.

Missing Context

In Taki Allen incident, school safety team reviewed AI alert, confirmed false positive, and canceled it before principal independently notified police.

Shows human verification occurred as intended, countering narrative of direct AI-to-police action without checks; principal's separate decision led to response.

Missing Context

Baltimore County Public Schools implemented Omnilert AI gun detection in 2023/2024 with human review protocols; Maryland Inspector General found no racial bias in Allen case.

Provides context on system's rollout and official review exonerating bias claim, balancing trauma focus.

Framing

Frames Taki Allen incident as "AI’s misidentification and the human decisions that followed" directly leading to trauma, omitting that school safety team verified and canceled the alert before principal's independent call to police.

Implies seamless AI-to-police action without human checks, exaggerating AI's direct role and downplaying safeguards.

Omission

No mention of AI gun detection system's successful detections or overall performance metrics.

Presents only false positives, creating one-sided view of risks without benefits or net impact.

Missing Context

Omnilert AI gun detection in Baltimore County schools has detected real threats, such as guns at Franklin Middle School, though not always the primary detection method in every case.

Balances the article's focus solely on false positives by showing the system has practical benefits in threat detection.

Missing Context

Studies and police reports indicate predictive policing can reduce crime by anticipating patterns, with LAPD Chief citing shifts to proactive deployment.

Provides counterbalance to the article's emphasis on risks without mentioning potential effectiveness or crime prevention benefits.

Framing

Presents Angela Lipps case as direct result of AI misidentification leading to 5 months jail, omitting that detention extended due to inability to post bail and other factors post-arrest.

Implies AI alone caused prolonged incarceration, overstating causal link without noting judicial/human decisions involved.

Source Credibility

One author, Steven L. Johnson, is co-chair for local Democrats and involved in progressive activism, not disclosed.

Potential undisclosed ideological incentive in critiquing policing tech.