WIREDGOV

Guest blog: Katie Taffler, UK Justice Sales Lead at Accenture as part of our #DigitalJustice2021 week.

Imagine a world where ……

• …… Prisons have a “smart” surveillance system that monitors every high-profile inmate in real-time via networked hidden cameras and sensors placed in every cell. As well as identifying suspicious behaviour from inmates, this close surveillance will also see an end to unethical behaviour from prison officers.
• ….. Prisons use video analytics to improve management of prison inmates, detecting abnormal behaviour, identify inmates entering an area they are not authorized to be in and provide accurate information on who inmates communicate with. This will provide prisons officers with a better ground situation awareness of activity to inform their interventions.
• …….Prisons use machine learning to look through data held on prison officers to identify and flag signs that an officer may be at a high risk of initiating an “adverse event”. This means systems become trained to identify officers that have a history of certain kinds of misconduct, or who may be under a lot of stress.

Sound futuristic? What if I told you these are just 3 of the many global digital advancements happening in the world right now –in China, Singapore and the US correspondingly. In the UK, we have a huge opportunity to learn from our global colleagues and capitalise on the momentum provided by Covid to meet one of the key performance objectives for HMPPS: Prison safety for both prison officials and prisoners themselves.

Two of the main threats to officer and prisoner safety I continually hear spoken about are (1) prisoner mental health issues, and (2) self-harm and the spread of extremism and organised crime groups (OCGs). And the latest figures demonstrate that the threat is rising; there has been a 135% increase in self-harm cases within prisons between 2010-2020, and a 75% increase in prisoners convicted of terrorism-related offences between 2015-2020. These are both issues which could be identified, intervened in, and prevented using the next generation of artificial intelligence (AI) technologies.

Creating ‘Smart Prisons’, using artificial intelligence via voice and video analytics, should be a key focus for the vision of Prison Reform. Voice analytics could be used to monitor calls that prisoners make to external parties, providing the prison service with significant insights into the behaviour and mindset of prisoners. Types of voice analytics such as speech recognition, sentiment analysis, transcription and translation capabilities would enable governors and officials to collate and scrutinize large amounts of call data – allowing a more personalised experience for each offender. Video analytics, which uses sophisticated computer algorithms to analyse CCTV footage against predefined rules, could also be used to alert officers to unusual patterns of behaviour. This footage can be used to detect Crowd Behaviour, monitoring prisoner populations in common areas and detecting prisoners who may be a risk to themselves or others. CCTV can also support Facial Recognition software, which would pinpoint vulnerable or threatening individuals within a populated area to help prison officers protect those in need and intervene on potential criminal or extremist activity. And finally, using an AI engine with prison CCTV could be used for Identification of Key Items to detect threatening items in prisoner possession, such as shanks, phones, drugs, and nooses, enabling prison officials to intervene and prevent a dangerous situation of self-harm or attacks on fellow inmates. None of this is particularly complicated of expensive technology – and indeed could sit on-top of existing estate CCTV hardware and phone terminals.

What’s more, these technologies are all already in place in parallel ‘high reliability’ organisations such as in Police and Revenue agencies in the UK, Ireland, France and Norway. Using AI to create the Smart Prisons of the future will enable HMPPS to better identify high risk individuals, intervene before a risk is realised, and prevent further danger to both officers and inmates – ultimately leading to a lower rates of recidivism post prison. The image most of us have of AI is silver robots, and the cultural hurdle is perhaps the first place to start. But with support from leadership, experienced technology partnerships, and user-centered design from the start, bringing AI into the everyday of prison activity can be a success we make in the prison system today – not one we wait until the next global pandemic to realise!

Katie Taffler

UK Justice Sales Lead; Accenture

With thanks to Jess Flannigan and Tom Whittaker for research and input via the Accenture Smart Prisons POV