fullscreen

eFinder

eFinder

Home › AI Ethics in Law Enforcement › Article

How AI can lead to false arrests and wrongful convictions

Phys · · 1369 words · By Maria Lungu; Steven L Johnson
AI Ethics in Law Enforcement Civil Liberties and Surveillance Probabilistic vs. Deterministic Systems
headphones Listen to the eFinder podcast briefing
Generate a natural audio summary of this story
Daily briefing

What to know about AI Ethics in Law Enforcement

The article discusses the risks of using artificial intelligence in policing, specifically how probabilistic outputs can be mistaken for certainties, leading to wrongful arrests. The authors argue for greater transparency regarding confidence thresholds and a better understanding of the difference between statistical likelihood and factual truth.

Propaganda risk 20%
Claims checked 9
Techniques found 1
Topics 3

Coverage spectrum

Coverage gap: Low Left coverage
Left0%
Center75%
Right25%

4 sources compared across this story cluster. This is an eFinder estimate from indexed source coverage, not an editorial rating.

What happened

How AI can lead to false arrests and wrongful convictions Gaby Clark Scientific Editor Andrew Zinin Lead Editor In Baltimore on Oct.

Why it matters

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.

Common ground

Within moments, police cars arrived, officers drew their weapons and Allen was forced to his knees and handcuffed while they searched him.

Perspective signals

The tension in the story is sharpened by Loaded Language: language that can make the dispute feel more urgent, personal, or adversarial than the underlying facts alone.

The article discusses the risks of using artificial intelligence in policing, specifically how probabilistic outputs can be mistaken for certainties, leading to wrongful arrests. The authors argue for greater transparency regarding confidence thresholds and a better understanding of the difference between statistical likelihood and factual truth.

open_in_new Read the original article: https://phys.org/news/2026-05-ai-false-wrongful-convictions.html

analyticsAnalysis

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

psychologyPropaganda Techniques Detected

eFinder identified 1 propaganda technique in this article. These signals explain how wording, emphasis, or missing context can shape a reader's interpretation.

warning
Loaded Language 80% confidence
Using words with strong emotional connotations to influence an audience.
Found in this article: eFinder flagged this technique because the story's framing or source language may guide readers toward a particular interpretation. Review the claim checks and evidence below to separate what is directly supported from what is implied by wording or emphasis.
Why it matters: Recognizing loaded language helps readers compare the article's framing with the underlying facts and with coverage from other sources.

fact_checkClaims Checked

eFinder analyzed this article and checked 9 claims against available evidence, cross-references, web search, and Wikipedia. Here is what the fact-checking layer found.

check_circle Corroborated 4
verified Verified By Reference 3
help Insufficient Evidence 2
verified
Claim 1: “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.”
VERIFIED BY REFERENCE
While search results mention the University of Virginia and general research on democracy/technology, no specific evidence was found regarding a 'Digital Technology for Democracy Lab' study on police chiefs' policies regarding human discretion versus AI.
menu_book
wikipedia NEUTRAL — Democracy is a form of government in which political power is vested in the people or the population of a state. Under a minimalist definition of democracy, rulers are elected through competitive elec…
https://en.wikipedia.org/wiki/Democracy
menu_book
wikipedia NEUTRAL — E-democracy (a blend of the terms electronic and democracy), also known as digital democracy or Internet democracy, uses information and communication technology (ICT) in political and governance proc…
https://en.wikipedia.org/wiki/E-democracy
menu_book
wikipedia NEUTRAL — The University of Virginia Center for Politics (CfP) is a nonpartisan institute at the University of Virginia. Based in Charlottesville, Virginia, the institute promotes the value of politics and the …
https://en.wikipedia.org/wiki/University_of_Virginia_Center_…
+ 3 more evidence sources
help
Claim 2: “AI systems rarely produce binary answers: yes or no, a positive identification or a negative one. They generate probabilities.”
INSUFFICIENT EVIDENCE
No evidence was provided in the search results to confirm or deny the specific claim about binary vs. probabilistic outputs of AI systems.
verified
Claim 3: “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 BY REFERENCE
Wikipedia entries on Generative AI and GPT explain that these models use deep learning architectures (transformers) to generate data based on patterns learned during pre-training, rather than retrieving specific facts from a database.
menu_book
wikipedia 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.wikipedia.org/wiki/Generative_AI
menu_book
wikipedia NEUTRAL — A generative pre-trained transformer (GPT) is a type of large language model (LLM) that is widely used in generative artificial intelligence chatbots. GPTs are based on a deep learning architecture ca…
https://en.wikipedia.org/wiki/Generative_pre-trained_transfo…
menu_book
wikipedia NEUTRAL — GPT-5.5 (Generative Pre-trained Transformer 5.5) is a large language model (LLM) released by OpenAI on April 23, 2026. The model is also known by its codename "Spud". OpenAI reports improvements on be…
https://en.wikipedia.org/wiki/GPT-5.5
+ 3 more evidence sources
check_circle
Claim 4: “The tools ingest historical crime data and score neighborhoods on predicted risk so officers can be routed toward the resulting hot spots.”
CORROBORATED
Multiple sources describe predictive policing as using historical crime data and algorithms to forecast 'hotspots' for resource allocation.
travel_explore
web search NEUTRAL — Efficiency in Resource Allocation: By analyzing historical crime data and predicting potential hotspots, law enforcement agencies can allocate resources more effectively using AI policing tools. This …
https://gsaschedule.ai/ai-policing-tools-for-law-enforcement…
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 — AI policing claims neutrality but instead amplifies existing racial bias, using opaque and unaccountable systems that disproportionately target and harm Black communities.
https://b1daily.com/2026/01/15/how-ai-policing-unfairly-targ…
check_circle
Claim 5: “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 event: Taki Allen, a student in Baltimore, was detained on October 20, 2025, after an AI system mistook a Doritos bag for a gun. Wikipedia confirms the existence of Kenwood High School in Baltimore County.
menu_book
wikipedia NEUTRAL — Kenwood High School is a Baltimore County public high school located in Essex, Maryland, United States.
https://en.wikipedia.org/wiki/Kenwood_High_School_(Maryland)
menu_book
wikipedia NEUTRAL — Comedy Bang! Bang! is a comedy podcast which began airing as a radio show on May 1, 2009. It is hosted by writer and comedian Scott Aukerman, best known for his work on the 1990s HBO sketch comedy pro…
https://en.wikipedia.org/wiki/Comedy_Bang!_Bang!
travel_explore
web search NEUTRAL — “I was just holding a Doritos bag – it was two hands and one finger out, and they said it looked like a gun,” Allen said. Baltimore county high schools last year began using a gun detection system usi…
https://www.theguardian.com/us-news/2025/oct/24/baltimore-st…
+ 2 more evidence sources
verified
Claim 6: “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 technical guides confirm that the Receiver Operating Characteristic (ROC) curve is used to illustrate the performance of a binary classifier 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…
help
Claim 7: “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.”
INSUFFICIENT EVIDENCE
No evidence was provided in the search results specifically discussing precision-recall analysis in the context of this claim.
check_circle
Claim 8: “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
Three independent web search results confirm that Angela Lipps of Tennessee was jailed for five months and released on December 24, 2025, due to a facial recognition error linking her to crimes in North Dakota.
menu_book
wikipedia NEUTRAL — A one-hit wonder is a musical artist who is successful with one hit song, but without a comparable subsequent hit. The term may also be applied to an artist who is remembered for only one hit despite …
https://en.wikipedia.org/wiki/List_of_one-hit_wonders_in_the…
menu_book
wikipedia NEUTRAL — The following deaths of notable individuals related to American television occurred in 2016.
https://en.wikipedia.org/wiki/2016_deaths_in_American_televi…
menu_book
wikipedia NEUTRAL — Clearview AI, Inc. is an American facial recognition company, providing software primarily to law enforcement and other government agencies. The company's algorithm matches faces to a database of more…
https://en.wikipedia.org/wiki/Clearview_AI
+ 3 more evidence sources
check_circle
Claim 9: “AI policing tools are used in dozens of U.S. cities, although no public registry tracks the full footprint.”
CORROBORATED
Multiple web search results confirm the adoption of predictive policing AI in various U.S. jurisdictions (e.g., Michigan) and note a lack of transparency and oversight/registries.
travel_explore
web search NEUTRAL — However, sheriffs and police chiefs tend to view AI assessments more favorably compared to local government officials and county prosecutors. A third of Michigan law enforcement leaders say their agen…
https://news.umich.edu/many-local-officials-in-michigan-unce…
travel_explore
web search NEUTRAL — Predictive policing AI, increasingly adopted by law enforcement, uses data analytics to forecast crime but raises significant ethical concerns, including privacy violations, lack of transparency, and …
https://techxplore.com/news/2025-05-policing-ai-accountable-…
travel_explore
web search NEUTRAL — Predictive policing. Police forces use AI tools to predict the risk of individual committing crimes. They can also predict the risk associated with certain areas. AI in football stadiums.
https://www.dailymail.com/sciencetech/article-13522765/AI-co…
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.