Using AI-Powered Immersive Technologies for Improved Situational Awareness of Police Officers: Promise and Challenges
A group of police officers is approaching the premises of a deserted factory, where they have information that a group of criminals are holding a victim of kidnapping. The officers are well-trained to handle these situations, but the criminals are known to be very dangerous, and they warn the officers that they have planted explosives in the premises over the last twenty hours. Luckily, one officer is equipped with augmented reality glasses, which utilise machine learning algorithms that have access to data from the CCTV network of the area, allowing them to receive real-time advice on the position of both the criminals and the explosives. The criminals are arrested, and the victim is safely rescued.
Does this sound like the scenario of a science-fiction movie? It is much closer to real-world police operations than one might think. Virtual and augmented reality (VR and AR) tools have been used for some time in other sectors for purposes of training personnel and improving awareness about the respective environment. For example, Nokia has developed a medical VR/AR technology and pilot-tested it in Dubai hospitals to enable doctors to get access to the bio-vital signs of patients while they are still in the ambulance and link them to their medical records to make a remote diagnosis and prepare immediate treatment more effectively. In the UK, the West Midlands Fire Service has been using a VR fire investigation training toolkit, developed by the UK-based company RiVR, to simulate fire investigation drills that will improve real-world responsiveness of their commanders.
Coming closer to the policing domain, a major barrier for immersive technologies had been the high levels of network bandwidth and processing capacity that VR and AR applications necessitate for the interpretation and representation of a complex crime scene. In the advent of machine learning and 5G technologies, however, these barriers can be overcome, and researchers have been developing cutting-edge immersive tools that can revolutionise real-time decision-making by police officers, much like the hypothetical scenario described above.
The project DARLENE, funded within the EU Horizon 2020 grant scheme from 2020 to 2023, is a prime example of this emerging research. DARLENE combines machine learning, AR and 5G technologies to allow police officers in the field to have an improved picture of what is happening, in other words to enhance their situational awareness. The current state of the art is that officers receive information about the ongoing incident and the characteristics of the operational environment from the police operations centre, often via wireless audio communication. Several difficulties arise in passing sufficient information from the operations centre to the field officer : it can be very difficult to aggregate information from multiple sources, i.e., police records, video feeds from local cameras, data belonging to private entities and other sources that could be relevant in assessing the likelihood of a crime occurring. Furthermore, the current means of communication are not tailored to the individual characteristics of police officers, e.g., their age and gender, even if it has been well-established that these factors are strongly correlated to the coping styles of officers in the field.
DARLENE aspires to produce an immersive technological solution, consisting in a head-mounted display unit, or in other words in ‘smart’ AR glasses. These glasses will use a wearable micro-computer module to perform demanding computer vision applications that will generate real-time information and advice that can be overlayed on the physical world and support officers to make accurate operational decisions in the most challenging of scenarios. At the same time, the physiological information of the officers will be fed back into the system, enabling the appropriate tailoring of information projected via the AR glasses based on the individual needs and stress tolerance. Through these functions, the technology could empower officers to identify friends or foes, vulnerable and injured individuals, violent behaviour, or suspicious unattended items in public spaces. It is hoped that this will allow law enforcement agencies to prevent criminal or terrorist activities and reduce the number of severe injuries or fatalities.
For all its promise, the technology also brings significant challenges related to its ethical, legal, and societal implications. Trilateral Research, as the Ethics Manager of the project, works to identify and mitigate potential risks emerging from the introduction of immersive technologies in policing. For example, we know that the quality of the information displayed is only as good as the data behind it. Considering existing biases in presently available violence datasets, where certain age and race groups tend to be overrepresented, an effort has been made together with technical partners to augment the datasets used for the training of the system to ensure a more balanced representation. Furthermore, it is important to design immersive technologies like DARLENE, so they do not ‘hyper-nudge’ police officers to certain decisions, effectively removing autonomous agency and moral responsibility behind their actions. As immersive technologies in policing become more prevalent, we should remember that we are developing these technologies to augment, not substitute human decision-making.
Author:
Stergios Aidinlis, Research Analyst, Trilateral Research
Georgie Morgan
Georgie joined techUK as the Justice and Emergency Services (JES) Programme Manager in March 2020, progressing to Head of Programme in January 2022.