The battlefield of the urban sprawl is changing, and innovative techniques in forensic science and computation are revolutionising police work. Police in the West Midlands are trial-running artificial intelligence technology funded by the EU and built by Middlesex University London called VALCRI, or “Visual Analytics for sense-making in Criminal Intelligence analysis”. It works by pouring through police records (interviews, crime scene photos, forensics) and spotting connections between cases. Currently, it is only being tested with data from three years of past cases, the suspects anonymised, with a human analyst working alongside it assessing the relevance of the spotted connections. Instead of painstakingly searching through a variety of evidence to find connections, VALCRI can collate and organise data to provide insights much faster than a single human analyst. Developments in machine learning could help to include the subtleties of cases, which would have caused problems for computers of the past. While the technology is promising, there are complications in the legal department. Professor Ifan Shepherd of Middlesex University London says “the data in a crime case is simply not good enough… A human analyst has to call the shots” (New Scientist). Police techniques can also be challenged in court, meaning a problems identified with a single VALCRI case could cause many other cases to be invalidated.
VALCRI is not the only new crime-fighting machine to recently go through a trial-run. In the U.S. capital, security robots are patrolling Washington Harbour centre, equipped with 360-degree cameras, thermal imaging, number plate recognition, and software designed to detect suspicious behaviour. Anything that is spotted is then reported to a human guard, who will take appropriate action. Despite a recent incident where one such robot “Steve” failed to pick up on a fountain in its path and fell into it back in July, robot security guards seem a promising technology. It could make busy areas safer by deterring criminals, despite the robots currently lacking the tools to “neutralize” a threat (which is probably a good thing for now; ED-209 from Robocop is an extreme possibility of malfunction leading to unfortunate accidents). Another benefit of robot guards is that they won’t tire from the long and potentially dull hours of patrolling. As Travis Deyle of Cobalt Robotics says, “security guard work is challenging because, mentally, very little is happening until it happens” (phys.org). The robots also can’t sleep on the job, and for an average of $7 an hour, they’re affordable too.
With all its benefits, there is still the question of how comfortable we are with handing over agency to technology that is still in development. Personally, I’d rather not be tasered to the ground because a robot profiled me as a “potential threat”. On the other hand, bias exists with any human police officer, the difference being a robot’s programming can be corrected much more easily than a human’s preconceptions. It is inevitable that we will start to see more robot technology in our day-to-day lives in the future but there is room for development before we start to rely heavily on it.