"In this economy AI can do basically anything, and if you haven’t noticed, it is. Why humans rely on AI so much these days is beyond me, but I do know that the police are being a little too trusting with AI, here’s what you need to know." -- Violet Regilio, 7th Grade
The Oakland police started using AI technology to solve crimes in 2012 but recently they started using AI-powered software in July 2025. However, the public has been worried about the mistakes made with these tools, which has led to bans on certain AI tools, like predictive policing and facial recognition.
People have been concerned about AI Facial Recognition because of several mistakes made by the tool. Route Fifty explains one of the “mistakes” facial recognition has made. “One story shared in the letter was that of a Black man in Georgia who was arrested for using stolen credit cards to buy designer purses in Louisiana. The man remained behind bars for six days before authorities realized his arrest was the result of a bad facial recognition match and that he had never been to the state.”
The New York Times also recognized this and stated “Mr. Reid’s wrongful arrest appears to be the result of a cascade of technologies — beginning with a bad facial recognition match — that are intended to make policing more effective and efficient but can also make it far too easy to apprehend the wrong person for a crime. “
So far Facial Recognition has a bad reputation, however the Medium says it’s a great tool. “Homeland Security Investigations (HSI) has also experimented with AI as a tool for cracking cold cases…. A retired HSI investigator, Jim Cole, has said the agency has been using similar technology to solve older cases. According to Cole, facial recognition has helped identify hundreds of victims and abusers.”
Still, many things can go wrong with AI and we shouldn’t trust it. “I have a few different worries about AI helping police,” said Paloma Bastos, an OSA Literary Arts student, “We have already had AI find peoples faces in the past and that did not go well. A few people got arrested for no reason, and it was mostly black people, too. They were treated horribly because AI thought that was the person who committed a horrible crime. It could also train itself to do things we have seen it do in the past, with the Facebook robots that created their own language and communicated inappropriately to each other, and they were AI.”
Other concerns people have with AI solving cases is predictive policing, an AI tool that predicts potential crimes before they happen. The article What’s Wrong with Predictive Policing? states “The most widespread criticism of place-based predictive policing is that it discriminates against people of color (Lum and Isaac 2016). This discriminatory treatment can result from a combination of two features of some predictive policing systems.” Duncan Purves the author of this article explains that this AI tool has a racial bias against people of color.
Another article titled Why ‘Predictive’ Policing Must be Banned states “Police forces across the UK are increasingly using so-called “predictive policing” technology to try to predict crime. Police claim these tools “help cut crime, allowing officers and resources to be deployed where they are most needed.” In reality, the tech is built on existing, flawed, police data… As a result, communities who have historically been most targeted by police are more likely to be identified as ‘at risk’ of future criminal behaviour. This leads to more racist policing and more surveillance, particularly for Black and racialised communities, lower income communities and migrant communities.”
Although many think that AI shouldn’t be allowed to be used in crime solving, in some cases AI has been a crucial tool in police departments. One case AI cracked was the famous 40 (plus) year old, Golden State Killer Case. They used genetic genealogy to solve this horrifying case.
The article 3 Examples of AI Helping Police Solve Cold Cases says “One of the most famous examples of AI aiding investigations is the cold case of the Golden State Killer. The killer, Joseph James DeAngelo Jr., pled guilty to 26 counts of murder in 2018 — decades after actually committing the crimes. Los Angeles police were able to find him after so long because of AI-assisted DNA analysis.” This shows AI can sometimes be helpful.
Another way AI is helping solve crimes in police departments is the AI tool gunshot detection. The Article How Cities Can Pair Audio Detection with AI-Powered License Plate Recognition and Video Technology to Revolutionize Addressing Gun Violence explains how this is a great tool and all cities should install them. “New AI-powered audio detection systems, like Flock Safety Raven, leverage machine learning to identify gunshots with over 90% accuracy. It provides real-time alerts within 60 seconds, pinpointing the incident location within 90 feet. This is significantly quicker than relying on citizens to call 911, ensuring that police can secure the scene, preserve evidence, and assist victims.” As we know shootings are more likely to happen these days so many think this AI tool is necessary.
Overall, there are many different opinions about AI. AI could help us in many ways, but we should have boundaries for how far it can go.
People have been concerned about AI Facial Recognition because of several mistakes made by the tool. Route Fifty explains one of the “mistakes” facial recognition has made. “One story shared in the letter was that of a Black man in Georgia who was arrested for using stolen credit cards to buy designer purses in Louisiana. The man remained behind bars for six days before authorities realized his arrest was the result of a bad facial recognition match and that he had never been to the state.”
The New York Times also recognized this and stated “Mr. Reid’s wrongful arrest appears to be the result of a cascade of technologies — beginning with a bad facial recognition match — that are intended to make policing more effective and efficient but can also make it far too easy to apprehend the wrong person for a crime. “
So far Facial Recognition has a bad reputation, however the Medium says it’s a great tool. “Homeland Security Investigations (HSI) has also experimented with AI as a tool for cracking cold cases…. A retired HSI investigator, Jim Cole, has said the agency has been using similar technology to solve older cases. According to Cole, facial recognition has helped identify hundreds of victims and abusers.”
Still, many things can go wrong with AI and we shouldn’t trust it. “I have a few different worries about AI helping police,” said Paloma Bastos, an OSA Literary Arts student, “We have already had AI find peoples faces in the past and that did not go well. A few people got arrested for no reason, and it was mostly black people, too. They were treated horribly because AI thought that was the person who committed a horrible crime. It could also train itself to do things we have seen it do in the past, with the Facebook robots that created their own language and communicated inappropriately to each other, and they were AI.”
Other concerns people have with AI solving cases is predictive policing, an AI tool that predicts potential crimes before they happen. The article What’s Wrong with Predictive Policing? states “The most widespread criticism of place-based predictive policing is that it discriminates against people of color (Lum and Isaac 2016). This discriminatory treatment can result from a combination of two features of some predictive policing systems.” Duncan Purves the author of this article explains that this AI tool has a racial bias against people of color.
Another article titled Why ‘Predictive’ Policing Must be Banned states “Police forces across the UK are increasingly using so-called “predictive policing” technology to try to predict crime. Police claim these tools “help cut crime, allowing officers and resources to be deployed where they are most needed.” In reality, the tech is built on existing, flawed, police data… As a result, communities who have historically been most targeted by police are more likely to be identified as ‘at risk’ of future criminal behaviour. This leads to more racist policing and more surveillance, particularly for Black and racialised communities, lower income communities and migrant communities.”
Although many think that AI shouldn’t be allowed to be used in crime solving, in some cases AI has been a crucial tool in police departments. One case AI cracked was the famous 40 (plus) year old, Golden State Killer Case. They used genetic genealogy to solve this horrifying case.
The article 3 Examples of AI Helping Police Solve Cold Cases says “One of the most famous examples of AI aiding investigations is the cold case of the Golden State Killer. The killer, Joseph James DeAngelo Jr., pled guilty to 26 counts of murder in 2018 — decades after actually committing the crimes. Los Angeles police were able to find him after so long because of AI-assisted DNA analysis.” This shows AI can sometimes be helpful.
Another way AI is helping solve crimes in police departments is the AI tool gunshot detection. The Article How Cities Can Pair Audio Detection with AI-Powered License Plate Recognition and Video Technology to Revolutionize Addressing Gun Violence explains how this is a great tool and all cities should install them. “New AI-powered audio detection systems, like Flock Safety Raven, leverage machine learning to identify gunshots with over 90% accuracy. It provides real-time alerts within 60 seconds, pinpointing the incident location within 90 feet. This is significantly quicker than relying on citizens to call 911, ensuring that police can secure the scene, preserve evidence, and assist victims.” As we know shootings are more likely to happen these days so many think this AI tool is necessary.
Overall, there are many different opinions about AI. AI could help us in many ways, but we should have boundaries for how far it can go.