How AI can lead to false arrests and wrongful convictions

In Baltimore County, Maryland, on October 20, 2025, a 17-year-old student named Taki Allen was mistakenly targeted by an AI-enhanced surveillance camera. The system falsely identified the Doritos bag in his pocket as a gun, prompting police to arrive with weapons drawn. Allen was forced to the ground, handcuffed, and searched before officers discovered only a crumpled bag of chips. Separately, in December 2025, Angela Lipps, a Tennessee grandmother, was arrested at gunpoint while babysitting her grandchildren. Police had used facial recognition software to incorrectly link her to fraud crimes in North Dakota, a state she had never visited. She spent five months in jail before being released after the error was uncovered. These cases highlight how AI systems, which operate on probabilities rather than certainties, can lead to severe consequences when treated as infallible. Researchers studying the intersection of technology, law, and public administration note that such tools are widely used in U.S. cities, often without full transparency or public oversight. AI policing tools analyze historical crime data to predict high-risk areas, guiding officers to deploy in those locations. However, these predictions—often presented as risk scores or heat maps—do not account for the uncertainty inherent in statistical models. Once a system flags a potential threat, law enforcement may act without questioning the confidence level behind the prediction. The issue extends beyond policing to other AI applications, such as generative models like ChatGPT or Claude. These systems generate responses based on patterns in training data, not verified facts. While they may produce statistically likely answers, they lack the context or accuracy of fact-checked information, creating risks when users assume the output is definitive truth. Experts warn that the shift from probabilistic predictions to operational certainty in law enforcement can have dangerous real-world consequences. Without proper safeguards, AI tools may deepen biases, misallocate resources, and lead to unjust outcomes for individuals.