Murder Predictive Policing Algorithm

Craig: so on minority report, they famously introduced precrines, where they had the precogs who would determine if you were gonna, commit a crime beforehand and you'd get arrested. and now it seems like the uk is using ai to do the same thing. they've got a murder prediction algorithm. what's happening here, john?

Jon: yeah. so first off, they're not using human beings or whatever those precogs were, like, in minority report. this is all, computer when i

Craig: say the same thing, i'm sorry. they were no. no. no.

Jon: i don't think you said it. i just wanna make sure that that we're clear. we're starting at everyone understands where we're starting from. so, yeah. so they're they're dubbing the homicide prediction project, and the tool uses machine learning algorithms to analyze datasets of a hundred thousand to 500,000 individual. i don't know that number. which includes, and this is scary, criminal records, mental health history, addiction, self harm, and domestic abuse reports. so the data sourced from the ministry of justice in the uk, the home office, greater manchester police, and the metropolitan police with inputs from suspects, victims, witnesses, and missing persons. so the scary part about this is that the data scope includes sensitive health markers like mental health, disability, vulnerability, and non conviction records. so suspects that weren't even charged. like, what so now in the uk, if your, record gets expunged or whatever, it's never expunged anymore, which is scary. this is this is dystopian scary crap. the other part of it, and don't even get me started, is on the bias. right? what are your thoughts, craig?

Craig: i'm always worried about this stuff, but, you know, we're looking at utilizing data analytics to forecast criminal activities based on historical records. so people talk about racial profiling. so they're not gonna so all of a sudden, they hand it off at ai, and they're just gonna call it something else?

Jon: they're gonna blame the ai for any problems.

Craig: yeah. and so it's not humans who are biased. it's the computer. so, you know, no one's in trouble for that. so i don't know. it feels a little bit like a scapegoat. but two, you know, isn't the uk in enough trouble right now with, you know, it seems like they're turmoil. they don't know which way they're going from a law point of view, and you're more likely to get arrested for making a meme than to do an actual crime. and now they wanna do precrime, so they're gonna guess that i'm gonna post a meme on facebook ahead of time, and they send the cops to my door and say, you were thinking about posting a meme. you're in trouble.

Jon: yeah. yeah. that that is one country i will never move to, based on what they're doing. this is crazy. it's nuts because there's so much privacy val violations. so they're collecting sensitive health data, without explicit consent, and that, i think, violates gdpr compliance issues. but, oh, they're the government, so they can probably, get around that. and that's i don't like that. i don't like what they're doing. i don't like this. i don't like the approach. there's so many issues that they're just diving right into this, and they're not talking about how they're going to, how they're going to address any of the issues. so, yeah. so, like, there's the data bias feedback loop. so predictive tools, they amplify historical policing biases. right? so if black neighborhoods are over policed, their crime data becomes overrepresented, which leads algorithms to recommend more patrols there. so it's kind of a self fulfilling prophecy of discrimination, this is just, yeah, this is just scary.

Craig: and then anybody at any time could put their thumb on the scales, i would imagine. and this is another thing when you don't have the checks and balances. and, you know, these predictive algorithms are bound by the quality of their data, and flawed historical records could compromise the reliability of predictions, perpetual existing biases, and policing, as you mentioned. but, the uk's implementation of murder prediction algorithm marks a transformative phase in crime prevention strategy. this proactive method seeks to identify individuals deemed high risk for violent behavior. however, challenges arise, including risk of misidentification of individuals and ethical ramifications of such technology, and critics highlight the lack of transparency regarding how data is used and the mechanics governing its application. so you feel like a person of interest, the phone rings and gives you a sin number, and you don't know whether they're gonna be the victim or criminal?

Jon: and so i really did a deep dive into this. okay? so i found some research in this. so, amnesty international did a report in 2025. they found that 33 uk police force forces use predictive tools that disproportionately target racialized and low income communities. over policing generates dirty data that created bias feedback loops. mit did a study on the victim data because, oh, okay. let's include victim data. that should help. no. even when using victim reports, instead of arrest data, predictive tools still skewed toward minority neighborhoods due to reporting disparities. wealthier white individuals report black suspects more often, which distort risk assessments. and finally, i'm gonna skip one that i did just because i'm gonna go to this one. there was us predictive policing failure. so programs in la and chicago were scrapped after ineffectiveness and racial bias. new york's strategic prevention unit reduced murders but 5% by 5%, but it faced criticism for targeting black and latino youth. so, like, there's red flags across the board with this. i i i think they're, they're going in the wrong direction with this, and they're not doing it cautiously, carefully. i think they're trying to rush to get something in, and, it's going to backfire. sorry. i'm passionate about this.

Craig: no. no. i agree with you. i think i sit back and say, well, what does backfire mean? it just seems like the government doesn't care.

Jon: well, thank you for raising that. so, again, with everything going on, it seems like there's no more accountability for anything that's happening. so, like, you're right. i think this is just gonna keep moving forward despite all of the errors, despite all of the protests, despite all of the problems because there's no longer accountability with anything anymore.

Craig: and it's scary. you know? again, every time there's a black box on what's happening and then we just agree with the outcome, buy stocks, you're gonna be a murderer, we're predicting an outbreak of disease over here, whatever. who's putting a thumb on the scale behind, and who's writing the ai prompt to get the output? you know? so when you start sitting back on this stuff, you gotta wonder that, is this just another tool that can be abused? now, i mean, in an ideal world, we want, to have personal accountability. we want criminals to, you know, face the repercussions of their actions, whatever they may be at the appropriate level. you know? you do a minor thing, minor stuff happens to you, whatever that might be. but we live in canada. it might be different in the states. it obviously is different in uk. but i want better tools for the police. yes. but i don't want overreach. i don't want your rights to be violated. i don't want people over patrolling a neighborhood because a computer told them to. so how do we prevent that?

Jon: we need the balance. and how do we get that balance? and there is nothing that i know of currently that we could use, that we could implement. there's no potential solutions that we could use to to mitigate this. something new needs to come along.

Craig: so this is just a reminder that the major majority port report is, real. it's not fiction, and the offender assessment system is live in the uk.

Jon: minority report.

Craig: what did i call it?

Jon: majority report.

Craig: majority report. well, maybe we're all criminals now. i don't know. the minority report is real, and the offender assessment system is live in the uk. this uk murder prediction algorithm bodies both innovation and inherent risks associated with predictive policing. as these technologies advance, deliberation on ethical and use of civil liberties remains imperative. so what do you think of this? is there a future in precrime, or does it scare the crap out of you? who was it that said show me the man, and i'll show you the crime? i

Jon: don't know.

Craig: and that, again, has to do with, being able to look backwards and find stuff. like, when all of a sudden, one day support to truckers, give them it is out of, justin trudeau's mouth a week later. it's the truckers are terrorists. we have to shut down their bank accounts. and if you followed my advice two weeks earlier and gave them money, you're now a terrorist. i'm like, what? what happened? yeah. so that kind of stuff is very scary, and that kind of comment was towards that. so let us know what you think. we'd love to hear your feedback.