MFIA Wins Disclosure of Connecticut School Choice Lottery Algorithm 

Rows of desks and chairs in a school classroom

This month, the Media Freedom and Information Access Clinic (MFIA) at Yale Law School won an order requiring the Connecticut State Department of Education to disclose the algorithm it uses to assign children to highly sought-after seats in Hartford-area magnet and technical schools. Following a two-day hearing, the Connecticut Freedom of Information Commission rejected the Education Department’s claim the algorithm is a “trade secret” that it can keep confidential. 

The case stems from the education department’s development of a computerized algorithm to decide which students are admitted to interdistrict magnet schools, which were created to reduce racial and socioeconomic isolation in Hartford-area public schools. Because there are far more applicants than openings in the magnet schools, the algorithm selects students for admission in a way designed to achieve diversity goals established in the settlement of a long-running discrimination lawsuit. But by declaring the algorithm to be its “trade secret,” the department prevented parents from knowing how their children’s school assignments had been made and the public from checking whether the algorithm assigned seats in a fair way, according to the clinic.

In April 2021, the MFIA Clinic helped Alicia Solow-Niederman, an Associate Professor of Law at the University of Iowa Law School, pursue several Freedom of Information Act requests regarding the school assignment process. Solow-Niederman sought to understand how automated decision-making systems were being used by Connecticut agencies and the extent to which state law ensured sufficient transparency and accountability of decisions being made by computerised algorithms. The education department denied much of her request for information by asserting that the algorithm used in its school choice lottery system was a “trade secret.” MFIA then filed a complaint with the commission. 

“We decided to pursue the case on behalf of Professor Solow-Niederman because the Freedom of Information Act exists to preclude such ‘black box’ governance,” said Kelsey Eberly, MFIA’s Floyd Abrams Fellow. “If government agencies can impose trade secret protection over the systems they develop to conduct government business, then results cannot be audited and the public cannot meaningfully debate the reasonableness of government actions.” 

The commission’s decision, issued June 7, directs the education department to disclose the inputs it uses in making school assignments and the weights given to these inputs. The commission’s order conveys that government agencies cannot assert trade secret status over the automated processes by which they carry out their statutory duties, at least when the process has no commercial application, such as the one to award seats at magnet schools. 

“Too often, agencies claim that the details surrounding the algorithms they use should be shielded from public oversight by trade secret protections,” said MFIA student Jonathan Gibson ’24, who worked on the case. “Going forward, the commission’s ruling makes clear that to claim trade secret protections, the government must show that competitive harm would result from disclosure.”

The State Department of Education has indicated its intention to appeal the commission’s order. 

The decision comes just days after MFIA celebrated another major milestone for algorithmic transparency in Connecticut, when the state legislature enacted a law requiring assessments and a public inventory of artificial intelligence and algorithmic tools in use by state agencies. 

The Media Freedom and Information Access Clinic at Yale Law School is a law school clinic dedicated to increasing government transparency, defending the essential work of news gatherers, and protecting freedom of expression by providing pro bono legal services, pursuing impact litigation, and developing policy initiatives. The clinic is a program of the Abrams Institute for Freedom of Expression and Information Society Project.