NYC algorithm task force makes broad suggestions in final report

A task force charged with studying New York City’s use of algorithms in executing government policies published its long-awaited report Wednesday after 18 months since being announced.

The 36-page report makes broad suggestions for how the city should manage the automated decision-making processes it uses to determine everything from crime prediction, building inspections, criminal-justice services placement of students in public schools. The task force leads by calling for the creation of a centralized organizational structure within City Hall to serve as a resource for guiding agencies’ management of their algorithmic decision-making systems. It also suggests adopting a “phased approach” to developing and institutionalizing any guidelines that central operation cooks up.

The report makes equally broad recommendations to “facilitate public education” about the use of algorithms in implementing city policy, including public reporting and engagement with “relevant communities” that might be disproportionately affected by automated decisions.

Mayor Bill de Blasio created the task force in May 2018 following a law passed by members of the New York City Council who had criticized that the city’s reliance on algorithmic decision-making resulted in racially biased policies. But the task force’s year-and-a-half of work did not always go smoothly. It was criticized for taking nearly a full year to hold its first public meetings, and some of its members said New York officials were slow to provide examples of systems they could analyze, despite lists of city-government algorithms being readily available from groups like New York University’s AI Now Institute.

The task force eventually held its first public meeting in April, which featured discussions of algorithms used by the New York Police Department. A second meeting was held in May on the topic of transparency.

But de Blasio appears to be acting quickly on the task force’s suggestion to create a centralized authority to oversee the city’s use of algorithms. On Wednesday, he ordered the creation of the new position of algorithms management and policy officer within the Mayor’s Office of Operations.

“I’m excited that the mayor is creating a position for someone who will carry forward the important work the city has done around algorithms and decision-making,” Jeff Thamkittikasem, the operations office’s director who also chaired the task force, said in a press release.

Still, not everyone who served on the task force was pleased with the report. The group acknowledged that it failed to find consensus on a number of issues, including how to best define how automated decision systems work, leading to at least one member to describe her experience serving on it as “a waste,” The Verge reported.