How AI and facial recognition are transforming police case management

Johann Hofmann, CEO of Griffeye, discusses why AI and facial recognition are more important than ever in law enforcement case management, and how the technology works in practice.

The concept of “machines who can think for themselves” has existed since the time of the Ancient Greek philosophers, however the first time the world actually saw AI in practice was in the late 1990s, when IBM’s Deep Blue technology became the first computer to beat a chess champion, Russian grandmaster Garry Kasparov. AI started coming into its own in the 2010s with the evolution of chatbots like Amazon’s Alexa and Apple’s Siri, as well as the use of AI to enable other technologies, such as facial recognition.

Johann Hofmann, CEO of Griffeye

Facial recognition came into mainstream use around the same time as AI because the combination of these technologies is what was required to make it properly usable. The concept of facial recognition was invented in the 1960s by Woodrow Wilson Bledsoe, who developed a system that could classify photos of faces using an electromagnetic stylus on coordinates of a grid to map facial features. As I’m sure you can imagine, this was a highly manual process, but it proved that this sort of technology would be viable. Facial recognition then started to crop up in places like airports for automated security screening, however many will remember it frequently failing at first, as the technology was adapting to public use. Today facial recognition is predominantly used by social media platforms to identify people in photos, in personal technology as a security feature (i.e. iPhone 7’s Face ID), at airport biometric gates and, of course, by law enforcement.

The common association between AI, facial recognition and law enforcement is that police use this technology to identify suspects (people on watchlists such as terrorists and gang members) in public places via CCTV or bodyworn cameras. However, there is also a hugely important part to play for this technology in case management as well which is often forgotten in public discourse, namely that it augments the human investigator who is sifting through ever growing volumes of criminal content in order to help prosecute criminals and protect victims.

In this article I’ll discuss why this new technology is important now more than ever in law enforcement case management, before going into an overview of how the tech works in practice, illustrating my argument with how it is used to fight real world crime such as child sexual abuse, gang crime and terrorism.

Content: one of the biggest challenges in case management

The evolution of modern technology has been both a blessing and a curse for law enforcement. Of course, new tech (including AI and facial recognition) has helped police to improve crime fighting efficiency, however the proliferation of technologies that produce content related to crime (from mobile phones to CCTV cameras) means that the burden in evidence management is heavier than it has ever been. In fact, some cases can contain up to 30TB worth of data (the equivalent of around 30 million images) which investigators have to trawl through in order to pick out evidence to identify criminals, victims and so on.

Knowing where to start and how to prioritise a case can be one of the hardest parts of an investigation and can leave officers feeling overwhelmed. As well, when starting from scratch it can be very challenging to cut through the masses of data to identify key details and visual attributes and as a result, evidence is often found too late, or not at all. One of the most critical things is that content may have been analysed and filed by another department fighting another case, which could offer up important links. However, if cases are managed manually and non-collaboratively, these links can often be missed, so deploying the right technology is critical.

The challenge is exacerbated by the increasingly intelligent tools used by criminals to create, store and share content as well. For example, in child sexual abuse cases, research launched in late 2019 identified that 79.2% of police said that content was being stored in the cloud, making it harder for investigators to trace users or the source of the content – Dropbox and Google Services are the platforms most frequently used. In the last few years, around a fifth (20.8%) of law enforcement have also seen an increase in the use of encryption or encrypted services to store and share illegal content. This also increases the confidence of criminals in producing this content, meaning that this combination of factors has seen monumental growth in the scale of data involved in criminal cases.

All in all, the huge volume of content (and increasingly innovative methods of producing and sharing it) slows down investigators, meaning that crimes take longer to solve, leaving victims at risk for longer and criminals on the streets.

Application: Artificial intelligence and facial recognition

Until recently, much of the analysis of content by investigators has been completed in a binary manner and to a large degree manually (meaning, they have to sift through images and videos and classify each one individually, which is a very long and stressful process). Furthermore, the lack of collaboration between departments has kept investigators working in silos. However, the introduction of technology such as AI and facial recognition (as well as tools that enable better collaboration) are transforming how evidence heavy cases are completed. In fact, half of child sexual abuse investigators said that a combination of technologies including AI, facial recognition, PhotoDNA and forensic tools help them fight cases faster.

In individual departments, these technologies can automatically classify and sort all content before an investigator sits down to start their work. The technology will flag content that is relevant to the case – whether this includes images that it can tell are explicit, contain known victims or criminals, or known details that relate to a crime scene. It will also group together similar images, meaning investigators are able to quickly look through all necessary and related evidence in order to make conclusions faster and in a more structured way than if they were having to comb through everything manually.

A lot of these critical details that link together or identify key content would be challenging for the human eye to detect as easily, especially an eye fatigued by analysing thousands of pieces of similar content, so the technology is absolutely critical in ensuring cases are completed as thoroughly as possible.

Alongside this technology, case management has recently been revolutionised by the development of tools which help forces to collaborate with each other. In the past (and in the present for many countries), forces worked in a completely siloed manner and would only collaborate if they knew another force was involved with a case for some reason. However, the introduction of tools such as the CAID (Child Abuse Image Database) that exists in the UK has created a nationwide database where law enforcement can cross check all images that come in with other forces to identify if anything has cropped up before that could link it to another case. The introduction of AI and facial recognition into a system like CAID will also enable forces to identify if people or objects appear in other cases as well.

Despite this technology going somewhat under the radar in the press, it has been truly transformative to the forces embracing it. New developments are seeing AI and the human brain merge, bringing about more exciting developments in how it will be able to be used. For example, the human brain is incredibly good at reacting to images it perceives as “shocking”, even if these images flash by at a very rapid pace. The AI can register when the human brain reacts to certain images using EEG, and group them together, improving the categorisation of evidence further and showcasing the value of humans working alongside machines in law enforcement. But of course, there is so much more that can be done as these tools evolve as well. For example, the tech is developing to also analyse text, which will be incredibly useful in the fight against terrorism.

The benefits of these technologies are massive and will only grow as it evolves. It helps investigators fight crimes faster, with less exposure to the material and the stress it causes, meaning crimes are fought faster with less of an impact on the investigator. I’m excited to see how, as the tech learns more and more and as humans develop more ways to innovate with it, law enforcement will be radically transformed, no doubt making the world a safer place.

Register today for IFSEC 2023

16-18 May 2023, ExCeL London | IFSEC 2023: Recognising the past, embracing the future

Join thousands of likeminded security and risk professionals at IFSEC 2023 in May, as the UK's largest and longest running security event looks ahead to what's next in the sector as it celebrates its 50th birthday. This year will see the launch of the IFSEC distributor network, while London's new Elizabeth Line makes travel to the venue easier than ever!

You’ll find hundreds of leading exhibitors from the physical and integrated security sector, showcasing all the latest in video surveillance, access control, intruder detection, perimeter protection and software solutions. Join the community and secure your ticket today!

Secure your FREE ticket today!