What happens when Norwegian-developed police technology meets one of Europe’s largest forensic communities?
Through two exercises in London, AI4Interviews and the London Metropolitan Police have tested how artificial intelligence, voice-controlled cameras and automated report support can contribute to better documentation, less follow-up work and more time for policing.When the London Metropolitan Police, responsible for around 10 million inhabitants and with approximately 1,400 forensic practitioners, set aside time, senior leadership and professional expertise to test Norwegian solutions, it sent a clear signal: AI4Interviews addresses a real need.That need is not primarily about technology. It is about better documentation, faster information flow, less duplication of work and more time for professional assessments.In January and April 2026, AI4Interviews and London Metropolitan Police Forensic Services carried out two joint exercises in London. The collaboration has been named LONO, short for London and Norway. The first exercise, LONO1, was based on realistic crime scene and investigative scenarios. The second, LONO2, moved the testing into the laboratory.The aim was not to put on a technology show. The aim was to find out what actually works when the police have to do their job under time pressure, with quality requirements, and with documentation that must withstand both professional and legal scrutiny.Four ways of working were testedDuring the exercises, British forensic practitioners worked on specific scenarios in which they had to document findings, assess evidence, share information and prioritise further work, just as they do in real cases.The difference was that different ways of working were compared:Jodapro, using the RealWear Navigator 520, was used for live streaming, dictation and documentation through a voice-controlled head-mounted camera. The forensic practitioner could describe what was being done while the work was in progress, without having to stop to write.The Capture app was used for structured logging, notes and case-related documentation along the way.AI support was used to transform raw material into structured draft reports. Not as final reports, but as first drafts that professionals could quality assure, correct and take responsibility for.Current practice, with manual notes and writing, was also used to provide a real basis for comparison.The difference in time spent was clear. In one of the tests, a forensic practitioner using Jodapro spent around ten minutes documenting crime scene assessments and observations. A colleague working with pen and paper was still writing an hour and a half later.Joe Marchesi, Operational Forensic Advisor at The Met, summed it up dryly on day two:– We are not writing today because now we have already learned.From free notes to structured draft reportsWhat particularly caught the interest of the London police was not only that speech could be turned into text. Several systems can do that. What was interesting was that spoken observations could be lifted into established templates and workflows.Alan Tribe, Director of Forensic Services at the London Metropolitan Police, saw significant potential:– The fact that the solution can be adapted to our own processes, and actually turn verbal notes into structured templates, is on a completely different level from what other generative AI methods can deliver today.He also emphasised that modernisation is not about artificial intelligence taking over the human role:– We are determined to reduce the time we spend on process and on remembering things. That is not the job of forensic experts. The job is to think, interpret and provide professional assessments.For the police, this is a crucial distinction. Artificial intelligence can support documentation, structuring and draft reports. But it is still the professional who assesses, corrects and stands behind the final conclusion.Why practise innovation?Practising innovation may seem unfamiliar. The police often train for acute incidents, emergency preparedness and crisis management. But new technology must also be tested realistically before it can be used safely and effectively.Kjeld Hendrik Helland-Hansen, a forensic practitioner in the Western Police District and a key contributor to the LONO collaboration, describes it this way:– Innovation is not about placing a shiny new app on top of a busy working day and hoping that people will be excited. It has to fit into the workflow. It has to withstand rain, gloves, stress, poor connectivity, time pressure, incomplete information and tired forensic practitioners who want to do the job properly.This was exactly what the London exercises were designed to examine. Not whether the technology looks good in a presentation, but whether it works when professionals are in the middle of their work.The exercises showed that forensic practitioners can quickly adopt new technology if it is relevant, practical and adapted to the workflow. Ruth Buckley, Forensic Advisor for Incident Examination at The Met and professional lead for the exercise, pointed out how quickly participants shifted their attention from the equipment to the work itself:– They had less than 45 minutes of training in Jodapro and were able to go out and concentrate on the crime scene work, not on how to use the tools. That is beyond expectations.That may be the best test of good technology: it should quickly become a natural part of the work.Not only for the biggest casesThe forensic practitioners from the London police were clear that this type of technology is not only relevant in the most serious cases.Crime scene examiner Craig Rose pointed to its value in everyday crime:– This is just as relevant in everyday crime as in major cases. At a burglary, I can be finished in around half an hour, everything is documented, and I am on my way to the next job.This is where the greatest potential lies. Not necessarily in individual cases, but in the volume: the many assignments, the many reports and the many hours currently spent on manual writing and follow-up work.When documentation can take place closer to the incident, while observations are still fresh, it can provide greater precision and reduce the risk of important information being lost. When draft reports can be structured more quickly, professionals can spend more time on the assessments that require human expertise.– When technology frees up time, it is not the technology that wins; it is the public, said Ruth Buckley, Forensic Advisor for Incident Examination at The Met, who also holds roles in several forensic organisations.The laboratory as a learning arenaLONO2 built on the experiences from January, but moved the testing into the laboratory environment at Lambeth HQ Directorate of Forensic Services in London.There, Jodapro, Axon bodycam, RealWear Arc 3, the Capture app, speech-to-text, automated report generation and new concepts for AI-supported quality control were tested in realistic laboratory processes.This is important because forensic work does not stop when the crime scene is left. Evidence must be examined, samples secured, findings described, assessments reviewed and reports written in a way that can withstand scrutiny.During LONO2, draft reports were generated based on laboratory work and dictation. The draft reports were marked as AI-generated and had to be quality assured by a professional.That is not a technical detail. It is a principle. In forensic work, it must always be clear who has professional responsibility for the final product.Rule of law as a requirement for innovationIn discussions about artificial intelligence, it is easy to focus mostly on efficiency. For the police, that is not enough. Technology must contribute to better quality, better auditability and stronger legal safeguards.That is why rule of law was a recurring theme in the collaboration with The Met. The technology was not intended to replace professionals, but to support them. It was not intended to remove control, but to make documentation better and more systematic.Buckley emphasised this during the exercise:– This is about better documentation and better processes in policing, which is absolutely fundamental to the rule of law.For AI4Interviews, this is important. The project is not only about efficiency, but about how new technology can be developed and used in a way that strengthens trust, quality and legal safeguards.What does the police gain from exercising together?The LONO collaboration shows that innovation in policing should not take place in isolation in a project room. It must be tested in dialogue with professionals, leadership, technology, law, research and practical constraints.When Norwegian solutions are tested in a large British professional environment, they are challenged by different procedures, different roles, different requirements and different ways of working. This creates valuable friction. The Met gets to test Norwegian tools in its own scenarios. Norwegian police gain insight into British methods, improvement suggestions and new ways of organising the work.Bente Skattør, project manager for AI4Interviews, believes this is one of the great values of the collaboration:– This was a win-win. There is enormous value in testing across national borders.She points out that the London exercises show that the needs are surprisingly similar:– The London exercise showed that this is not a niche idea, but that we have solutions that address a global need: better documentation without more people. This is not the future. It is practice. And it works.The way forwardLONO1 and LONO2 are not the end of the work. The exercises are part of a longer development process in which experiences will be documented, analysed and used further.AI4Interviews and The Met are planning further professional collaboration, sharing of experiences and documentation of findings from the exercises. The aim is to further develop the solutions based on concrete feedback from the people who will actually use the technology.Helland-Hansen believes this is how innovation in policing should take place:– Not one demonstration, not a pilot study that dies in a drawer, but repeated testing, adjustments, documentation, critical questions and collaboration with professionals who actually know the everyday reality.For the police, this is ultimately about the core task: delivering good policing for the public. If technology can provide better documentation, reduce duplication of work and free up time, the gain can be significant. Not only for the police, but for the public.LONO is a collaborative project between London Metropolitan Police Forensic Services and AI4Interviews, with participants from Oslo Police District, Western Police District and the Norwegian Police IT Unit.Although forensic work has been the main theme in London, the experiences point far beyond one professional field. The principles can be transferred to the entire criminal justice chain: from the first incident to completed investigation and decided case.Artificial intelligence can help make policing more efficient. But the London exercises also show something more important: technology must be developed together with professionals, tested in realistic situations and used in a way that strengthens quality, responsibility and the rule of law.Read more about AI4Interviews
