Governance Under Threat? AI in the Social Rehabilitation System of Ecuador: An Application of the Bayesian Model

Measuring the interaction between the prison population, including violence, intramural and extramural insecurity, and misconduct can be a complex task, as there are several factors to consider. Some key variables we will measure are: intra-prison violence, extramural violence, misconduct, population demographics, staffing, and prison infrastructure. Overall, measuring these variables can help identify patterns and trends of violence and misconduct within the prison population, which can inform the development of effective prevention and intervention strategies. The development of a Bayesian model requires specifying the conditional relationships between the variables identified in the above analysis. In the context of the prison population, the Bayesian model would aim to predict the probability of violent acts or misconduct from the observed values of the identified variables. Numerical values could be assigned to the variables, such as frequency counts or percentages, and used to calculate conditional probabilities in the model. Finally, an artificial intelligence-based application is proposed to support decision-making within the Social Rehabilitation System of Ecuador.