## Article Summary Essex Police is currently under scrutiny for its implementation of facial recognition technology, which has raised significant concerns regarding its effectiveness and legality. The police force has reportedly scanned the faces of around 2.5 million individuals using a system called Apollo 4, developed by the Israeli company Corside. This technology is designed to compare faces in real-time against a watch list of individuals of interest, generating a confidence score to determine potential matches. However, the system's threshold for matching was set at 0.6, based on research from a different system used by the Metropolitan Police, leading to a fundamentally flawed calibration that risks flagging innocent individuals as suspects. The lack of formal testing prior to deployment has further exacerbated the situation. A Freedom of Information request revealed that no validation was conducted before the system was rolled out to the public. A subsequent study by Cambridge University found that the system only correctly identified 51% of targets, effectively making its accuracy akin to a coin toss. Additionally, the study highlighted evidence of ethnic bias and concluded that the system had no measurable impact on crime reduction, raising serious questions about the efficacy of such surveillance technologies. The implications of these findings are profound, particularly concerning data protection and privacy rights. Big Brother Watch has described the use of live facial recognition for everyday policing as authoritarian, inaccurate, and ineffective. The deployment of this technology without proper oversight raises significant legal concerns, especially regarding the processing of biometric data under the Data Protection Act of 2018 and UK GDPR. If individuals are wrongly identified as suspects, it could lead to unlawful processing of their personal data, further complicating the legal landscape surrounding such surveillance practices. As a result of these issues, individuals who may have been scanned and flagged by the system are encouraged to consider filing subject access requests to understand what data has been processed about them. The situation serves as a stark reminder of the importance of proper implementation and oversight of surveillance technologies by law enforcement agencies. While the police play a crucial role in maintaining public safety, the deployment of such systems must be conducted with the utmost care to avoid potential legal repercussions and to protect citizens' rights. The ongoing discourse surrounding Essex Police's facial recognition technology highlights the need for a balanced approach to policing that respects individual privacy while ensuring public safety. ## Full Transcript ## Essex Police and Facial Recognition Technology Essex Police may be facing serious trouble, and I suspect they won't be the only force in this situation. I predicted this might arise because, as I mentioned before, when a police force or organisation collects large amounts of data, they must set up and use those systems properly. Otherwise, it could lead to significant issues. Before we proceed, please subscribe to the channel. We are racing towards 1 million subscribers, and I am incredibly grateful to you all. Google informed me this week that you are a fantastic audience, engaging with my videos through thousands of comments and likes. So, please keep it up, and I will be very appreciative. ## The Issues with Essex Police's Facial Recognition System Now, what are we discussing? Essex Police has reportedly scanned the faces of approximately 2.5 million people using their facial recognition technology as it has been rolled out. You may have heard my complaints about this in the past, and I will revisit some of those points shortly. It turns out they set up the system incorrectly from the beginning—not just slightly wrong, but fundamentally flawed. They deployed a system called Apollo 4, developed by an Israeli biometrics company called Corside. This system scans faces in real-time and compares them against a watch list of predetermined individuals of interest. It produces what they call a confidence score, indicating whether your face matches someone on this list. This score determines whether you are flagged as a potential match and, therefore, of interest while you are simply going about your business. To achieve this, they need to set a threshold for matching. If the threshold is too low, it results in false positives, flagging innocent people as suspects. Conversely, if it is set too high, they may miss actual suspects. Essex Police reportedly set this threshold at 0.6, but they justified this by relying on research from an entirely different system used by the Metropolitan Police, which does not even use the same scale. Essentially, they calibrated one system using settings designed for another, akin to tuning a basic car with the algorithms of a super sports car. This is not just a grey area of overlap; it is a fundamentally flawed implementation of a system capturing biometric data. ## Lack of Testing and Serious Concerns According to a Freedom of Information request uncovered by Big Brother Watch, no formal testing was carried out. They did not validate the system, meaning they deployed it on the public without proper trials. The situation worsens: after almost a year of operation, Cambridge University tested the system and found that it correctly identified just 51% of targets. This means it was effectively a coin toss—missing half of the people it was supposed to identify. The study also revealed evidence of ethnic bias, which contributed to the decision to pause the system. Most importantly, the Cambridge study found that the system had no measurable impact on crime reduction whatsoever, neither during its deployment nor afterwards. In essence, over the course of a year, they scanned 2.5 million people's faces for biometric data without proper setup or testing, missed over half of the potential targets, and delivered no actual benefit in terms of crime reduction. This raises serious questions about the effectiveness of surveillance technology, which is often touted as a means to enhance safety. ## Questions of Law and Data Protection Big Brother Watch bluntly stated, "Live facial recognition for everyday policing is authoritarian, inaccurate, and ineffective." Given the evidence from this study, it is difficult to argue against that assertion. This situation raises significant questions about the legality and protection of data, as it involved real people and real surveillance. If this is the standard for how these systems are used, there are serious concerns about whether this technology intrudes on privacy and whether there are potential claims for breaches of data protection. Capturing biometric data through facial recognition scans and matching it against a database constitutes processing of that data. If someone is wrongly identified as a suspect, it could lead to unlawful processing of personal data under the Data Protection Act of 2018 and UK GDPR. The data must be collected properly, fairly, and accurately, and it must be proportionate to achieving a legitimate aim. While they may argue that the legitimate interest is to catch individuals in specific areas, there are still questions about whether the data is genuinely deleted after use. If someone has been wrongly flagged, this poses a serious problem. ## Potential Claims Against the Police If you have been questioned by the police in the Essex area over the past year, you may want to consider whether you were scanned and flagged by this system. You might file a subject access request for all data they may have processed about you in that area. I would be very interested to hear from anyone who has been misidentified or wrongly detained as a result of this system. I want to emphasise that I respect the police; they have a difficult job. However, when deploying these systems, they must do so properly. Otherwise, they risk criticism and potential legal action. Let me know your thoughts in the comments, and as always, thank you for watching.