Imagine you are part of the city government responsible for urban planning. Write a letter to address the biggest ethical obligation behind a LASER-like system.
As a division in charge of urban planning for the City of Los Angeles, we are obligated to ensure that the LASER system delivers sustainable and equitable improvements for our neighborhoods, especially the socially marginalized ones. To that end, we need LASER to generate outputs that would not target racial groups – African American communities, for example – and in turn aggravate the pre-existing police harassment in those regions. Yet, there exists no possible source of data inputs for LASER to achieve such goal. The algorithm has been historically accused of racial profiling despite not using race as a predictor for criminal tendencies. However, our society is fraught with deep-rooted racism such that data points themselves bear the mark of discriminatory ideology and practices. For instance, gang membership as a predictor in LASER essentially serves as a proxy for race. Gangs in African American communities arose as a result of economic deprivation, resource insufficiency, as well as police harassment – phenomena with roots that can be  easily traced back to age-old racial segregation. Similarly, criminal records hold the vestiges of the New Jim Crow. Therefore, using gang activity and past arrests as predictors in the LASER system only generate outputs with heavy racial connotations, which in turn reinforce and reify historical biases.

Notwithstanding the problematic algorithm, we do not need to cast away the system entirely. When LASER points to places of deep-rooted discrimination and deprivation, we should mend but not aggravate. Instead of composing chronic offender bulletins, we can treat the LASER outputs as agenda for our next social welfare and infrastructure construction projects – where do we need to build more schools and hospitals? Where should the money go to create more job opportunities? Where do we need to address the problem of police misconduct and rebuild the relationship between people and law enforcement? If we cannot fix a “bias in, bias out” algorithm, we can at least attend to the information it has generated and figure out where the biases have wreaked havoc and taken a heavy toll. We can then redistribute resources and fundings accordingly so that the historically deprived neighborhoods have the chance to flourish. In doing so, we can achieve both sustainability and equity. We can expect the crime rates to go down when the neighborhood achieves economic growth and employment, and such crime reduction is sustainable, unlike the original LASER system that continues to discriminate and deprive.