The Uncertainties of Risk Assessment

The hard and complex work of implementing new multi-agency risk assessment approaches to policing domestic abuse

“…First, the disjuncture between assessing risk, as required within any MARAC process, but doing so whilst using a holistic, trauma-informed approach. Risk is a structurally neutral concept which does not account for the ways in which intersectional constraints (such as gender and ethnicity) can impact behaviour (Mythen 2014 ), therefore adopting a risk-focussed approach whilst attempting to be trauma-informed creates a fundamental tension (see inter alia Hannah-Moffatt 2015 ). Secondly, understanding which cases are considered as part of the MARAC process, with a recognition that many cases of DA may not be assessed as ‘high risk’ using tools such as the DASH, yet may feature highly risky behaviours constituting coercive control (Barlow and Walklate 2021 ; Walklate et al 2021 ).…”

Risk and punishment: The recent history and uncertain future of actuarial, algorithmic, and “evidence‐based” penal techniques

“…Further, some have challenged the idea that risk tools augment transparency, accountability, and objectivity (see Hannah-Moffat, 2015). In some cases, the algorithms behind risk assessment tools are not publicly available for scrutiny.…”

Editorial: Applications of artificial intelligence in forensic mental health: opportunities and challenges

“…In addition, AI has been used to inform decisions about sentencing, parole, probation or pretrial risk assessment, raising several legal and ethical concerns regarding, among others, fairness, accountability and transparency ( Tortora et al, 2020 ). These concerns stem, for instance, from the finding that some algorithms contained race and gender bias ( Barabas et al, 2018 ), the fact that the scores may be misapplied and misinterpreted by judges and practitioners ( Hannah-Moffat, 2015 ), and proprietary issues which can contribute to a lack of transparency ( Barabas et al., 2018 ).…”

“…In addition, AI has been used to inform decisions about sentencing, parole, probation or pretrial risk assessment, raising several legal and ethical concerns regarding, among others, fairness, accountability and transparency (Tortora et al, 2020). These concerns stem, for instance, from the finding that some algorithms contained race and gender bias (Barabas et al, 2018), the fact that the scores may be misapplied and misinterpreted by judges and practitioners (Hannah-Moffat, 2015), and proprietary issues which can contribute to a lack of transparency (Barabas et al, 2018).This Research Topic aims to present cutting-edge studies on the application of AI techniques in the forensic mental health field, including research on ethical challenges such as those related to the need to ensure non-discrimination, the "fair process" and the values of transparency and comprehensibility of decision-making processes.…”