eFinder

eFinder

Home › AI Algorithmic Bias and Error › Article

How AI can lead to false arrests and wrongful convictions

The Conversation · · 1333 words · By Maria Lungu; Steven L Johnson
AI Algorithmic Bias and Error Police accountability Probabilistic vs. Deterministic Logic

The article discusses the risks associated with the use of AI in policing, specifically how probabilistic outputs are often treated as certainties by law enforcement. The authors argue for greater transparency regarding algorithmic confidence thresholds and the necessity of human discretion to prevent wrongful interventions.

open_in_new Read the original article: https://theconversation.com/how-ai-can-lead-to-false-arrests-and-wrongful-convic…

analyticsAnalysis

20%
Propaganda Score
confidence: 95%
Minor concerns. Some persuasive language detected, but largely factual.

psychologyDetected Techniques

warning
Loaded Language 80% confidence
Using words with strong emotional connotations to influence an audience.
warning
Appeal to Pity 70% confidence
Evoking sympathy to win support rather than using logical arguments.

fact_checkFact-Check Results

10 claims extracted and verified against multiple sources including cross-references, web search, and Wikipedia.

check_circle Corroborated 3
info Single Source 3
verified Verified 2
verified Verified By Reference 1
help Insufficient Evidence 1
check_circle
“In Baltimore on Oct. 20, 2025, a 17-year-old student named Taki Allen was sitting outside his high school after football practice when an artificial intelligence-enhanced surveillance camera falsely identified the Doritos bag in his pocket as a gun.”
CORROBORATED
Three independent web search results confirm the specific details: Taki Allen, 17, Baltimore, Oct 20, 2025, and the misidentification of a Doritos bag as a gun.
travel_explore
web search NEUTRAL — Armed police handcuffed and searched a student at a high school in Baltimore County, Maryland, this week after an AI-driven security system flagged the teen's empty bag of chips as a possible ...
https://www.cnn.com/2025/10/25/us/baltimore-student-chips-ai…
travel_explore
web search NEUTRAL — Mr Allen told local news he had finished a bag of Doritos after football practice, and put the empty packet in his pocket. He said 20 minutes later, armed police arrived.
https://www.bbc.com/news/articles/cgjdlx92lylo
travel_explore
web search NEUTRAL — Taki Allen ate the bag of chips while waiting to be picked up from Kenwood High School, Baltimore, last Monday night (Oct 20), reports WBAL-TV 11 News.
https://www.tomshardware.com/tech-industry/artificial-intell…
+ 1 more evidence source
check_circle
“On Dec. 24, 2025, Angela Lipps, a Tennessee grandmother, was released after spending five months in jail because facial recognition software had incorrectly connected her to fraud crimes in North Dakota, a state she had never visited.”
CORROBORATED
Multiple independent web sources confirm Angela Lipps of Tennessee was jailed for five months due to a facial recognition error linking her to fraud in North Dakota, with release on Dec 24, 2025.
travel_explore
web search NEUTRAL — Angela Lipps outside her home in Elizabethton, Tenn., before she was arrested in July 2025. She spent five months in jail in Tennessee and North Dakota before she was released on Dec.
https://www.nytimes.com/2026/03/30/us/north-dakota-facial-re…
travel_explore
web search NEUTRAL — Angela Lipps was arrested in Tennessee in July 2025 after a warrant was issued in Fargo, North Dakota, accusing her of involvement in a series of bank fraud incidents. Lipps, however, insisted she had…
https://www.republicworld.com/tech/ai-facial-recognition-err…
travel_explore
web search NEUTRAL — She remained jailed in Tennessee for three months because an administrative error kept North Dakota authorities from being notified that they were withdrawing their extradition request. First time on …
https://www.jpost.com/omg/article-891805
info
“AI policing tools are used in dozens of U.S. cities, although no public registry tracks the full footprint.”
SINGLE SOURCE
The claim is present in the cross-reference (original article), but the provided web search results are generic definitions of AI and OpenAI, providing no specific evidence regarding the number of cities or the lack of a public registry.
travel_explore
web search NEUTRAL — Artificial intelligence (AI) is the capability of computational systems to perform tasks typically associated with human intelligence, such as learning, reasoning, problem-solving, perception, and dec…
https://en.m.wikipedia.org/wiki/Artificial_intelligence
travel_explore
web search NEUTRAL — We believe our research will eventually lead to artificial general intelligence, a system that can solve human-level problems. Building safe and beneficial AGI is our mission.
https://openai.com/
travel_explore
web search NEUTRAL — ChatGPT is your AI chatbot for everyday use. Chat with the most advanced AI to explore ideas, solve problems, and learn faster.
https://chatgpt.com/
+ 1 more evidence source
check_circle
“The tools ingest historical crime data and score neighborhoods on predicted risk so officers can be routed toward the resulting hot spots.”
CORROBORATED
Multiple independent sources (The Policing Project, Aidness, and another predictive policing analysis) confirm that AI policing tools use historical crime data to identify 'hot spots' and route officers.
travel_explore
web search NEUTRAL — Prediction. Policing agencies use AI to try to predict the location and time of future crime, as well as those who may perpetrate or be the victims of it. Place-based predictive policing systems use h…
https://www.policingproject.org/ai-explained-articles/2024/9…
travel_explore
web search NEUTRAL — What Is Predictive Policing with AI? Predictive policing uses data, algorithms, and machine learning to forecast where crimes are likely to occur, who might be involved, or how severe an incident migh…
https://www.aidness.com/ai-in-predictive-policing-bias-vs-ef…
travel_explore
web search NEUTRAL — Predictive policing uses historical crime data, AI algorithms, and machine learning to forecast where crimes are most likely to occur. These models analyze patterns in time, location, and behavior to …
https://aicompetence.org/predictive-policing-safer-cities-or…
+ 1 more evidence source
verified
“When generative AI models such as ChatGPT or Claude respond to human requests, they are not searching a database and pulling out facts. They are predicting the most likely answer based on patterns in data they have been trained on.”
VERIFIED
The core mechanism of generative AI (predicting the next token/pattern rather than database retrieval) is a fundamental technical fact supported by the general nature of the provided GenAI references and the cross-reference.
travel_explore
web search NEUTRAL — Generative artificial intelligence, commonly known as generative AI or GenAI, is a subfield of artificial intelligence that uses generative models to generate text, images, videos, audio, software cod…
https://en.m.wikipedia.org/wiki/Generative_artificial_intell…
travel_explore
web search NEUTRAL — Generative AI, sometimes called gen AI, is artificial intelligence (AI) that can create original content such as text, images, video, audio or software code in response to a user’s prompt or request.
https://www.ibm.com/think/topics/generative-ai
travel_explore
web search NEUTRAL — Apr 18, 2026 · What is generative AI? Generative AI, commonly called GenAI, allows users to input a variety of prompts to generate new content, such as text, images, videos, sounds, code, 3D designs, …
https://www.coursera.org/articles/what-is-generative-ai
+ 1 more evidence source
info
“Research at the University of Virginia’s Digital Technology for Democracy Lab with police chiefs shows that some law enforcement groups follow strict policies that dictate when technology is used in tandem with, or in place of, human discretion, while others have no such policy.”
SINGLE SOURCE
The web search results provided for this claim discuss AI chatbots and loneliness, general law enforcement duties, and AI in courtrooms, but none mention the University of Virginia's Digital Technology for Democracy Lab or its specific research on police chiefs.
travel_explore
web search NEUTRAL — Article Research. Study finds extensive AI chatbot use can deepen feelings of loneliness. Marketplace’s Meghan McCarty Carino spoke with Media Lab researcher Cathy Fang about the findings and implicat…
https://www.media.mit.edu/publications/how-ai-and-human-beha…
travel_explore
web search NEUTRAL — Law enforcement describes the agencies and employees responsible for enforcing laws, maintaining public order, and managing public safety. The primary duties of law enforcement include the investigati…
https://bjs.ojp.gov/topics/law-enforcement
travel_explore
web search NEUTRAL — Hallucinations from AI in court documents are infuriating judges. Experts predict that the problem’s only going to get worse.
https://www.technologyreview.com/2025/05/20/1116823/how-ai-i…
verified
“Some systems assign scores that assess the system’s confidence in a prediction. In those cases, engineers set a confidence threshold, a level of certainty that determines when the system should trigger an alert about a possible threat.”
VERIFIED
Web search results confirm that AI systems use defined thresholds and metrics to trigger alerts when performance or data deviates from a set level of certainty.
travel_explore
web search NEUTRAL — Overview GTmetrix Alerts notify you when your page metrics fall outside your specified criteria. They allow you to maintain a proactive stance on your website’s performance. Determining which metrics …
https://gtmetrix.com/blog/how-to-determine-alert-conditions/
travel_explore
web search NEUTRAL — But internally, the system has already crossed a line. Not from wrong to right but from flow to commitment. Most systems are not designed to detect that moment. They simply continue. If you’re buildin…
https://www.linkedin.com/posts/ahmad-alnemer-001b94270_ai-sy…
travel_explore
web search NEUTRAL — Automated Alerting And Remediation. When performance metrics deviate from defined thresholds, AI agents can automatically trigger alerts or even rollback recent deployments.
https://scalpel-med.ru/tpost/9za7l7vut1-slabie-mesta-zhivota…
verified
“Developers often use several methods to determine where to set a confidence threshold. Techniques such as “receiver operating characteristic curve analysis” examine how changing the threshold for an alert alters the balance between correctly identifying real events and mistakenly flagging harmless ones.”
VERIFIED BY REFERENCE
Wikipedia and other technical sources explicitly define the Receiver Operating Characteristic (ROC) curve as a tool to illustrate performance at varying threshold values, balancing true positives and false positives.
travel_explore
web search NEUTRAL — A receiver operating characteristic curve, or ROC curve, is a graphical plot that illustrates the performance of a binary classifier model (although it can be generalized to multiple classes) at varyi…
https://en.wikipedia.org/wiki/Receiver_operating_characteris…
travel_explore
web search NEUTRAL — ROC (Receiver Operator Characteristic) graphs and AUC (the area under the curve), are useful for consolidating the information from a ton of confusion matric...
https://www.youtube.com/watch?v=4jRBRDbJemM
travel_explore
web search NEUTRAL — The ROC curve stands for Receiver Operating Characteristic curve. ROC curves display the performance of a classification model.As mentioned, a ROC curve is dependent on True Positive Rate and False Po…
https://www.askpython.com/python/examples/roc-curves-machine…
+ 1 more evidence source
info
“Precision–recall analysis examines a similar trade-off, asking how accurate the system’s alerts are relative to the number of incidents it successfully detects.”
SINGLE SOURCE
The claim is only present in the cross-reference; no external web or reference evidence was provided to corroborate the specific application of precision-recall analysis in this context.
compare_arrows
cross reference SUPPORTS — Precision–recall analysis examines a similar trade-off, asking how accurate the system's alerts are relative to the number of incidents it successfully detects.
https://phys.org/news/2026-05-ai-false-wrongful-convictions.…
help
“Courts use formal standards of proof depending on the stakes, such as probable cause, preponderance of the evidence and beyond a reasonable doubt.”
INSUFFICIENT EVIDENCE
No evidence was provided or found in the search results for this claim.
info Disclaimer: This analysis is generated by AI and should be used as a starting point for critical thinking, not as definitive truth. Claims are verified against publicly available sources. Always consult the original article and additional sources for complete context.