Some of the most controversial technologies government agencies use to surveil the public or automate decisions about them are developed or overseen by private parties.
Whether it’s automated license plate readers (ALPRs), cell-site simulators, or algorithmic tools used by federal courts and other agencies to make decisions about people’s life and liberty, the federal government increasingly purchases the technology from private contractors.
But a Freedom of Information Act (FOIA) case before the U.S. Supreme Court threatens to further restrict the public’s ability to learn about when government uses technology. That’s why EFF, along with a coalition of other organizations, filed a friend-of-the-court brief on Monday asking the court to decline an effort to expand a FOIA exemption that prohibits the disclosure of private parties’ trade secrets and confidential business information.
FOIA must permit the public to access to this information, the brief argues, because “private-sector technologies increasingly define how government programs operate, how they affect individuals, and whether they may infringe on constitutional rights and liberties.”
The underlying case, Food Marketing Institute v. Argus Leader Media, concerns a FOIA request by a newspaper that sought records on government funds paid to grocery stores and other retailers as part of the Supplemental Nutrition Assistance Program (SNAP), or what was formerly called the food-stamp program. The reporter was interested in seeing records about potential retailers defrauding the program.
The U.S. Department of Agriculture refused to release the information, claiming it was exempt under FOIA’s Exemption 4, which allows an agency to withhold information that contains trade secrets or confidential business information obtained from third parties, such as companies. A federal court found that the information requested did not qualify as confidential business information and ordered the agency to disclose it. The Food Marketing Institute, a trade group for grocery stores, intervened in the case and ultimately petitioned the Supreme Court to review that ruling.
Although the case is ostensibly about access to SNAP records, EFF and others are worried about the consequences should the Supreme Court adopt a broad reading of what qualifies as confidential business information under FOIA Exemption Four. As the brief argues:
An expansion of Exemption 4 would be particularly devastating for the public’s ability to understand government programs that increasingly depend on emerging and complex technology developed by private companies. The government relies extensively on the private sector to provide technology that is central to all manner of government activities—from “big data” algorithmic decisionmaking systems, to powerful surveillance technology, to the government’s core information infrastructures—and this reliance is sure to increase going forward. It is critical to cabin Exemption 4’s reach so that the public is not left without the ability to understand core governmental activities by accessing records about the private sector technologies on which those activities depend.
The brief highlights how public access to information about controversial government programs, including facial recognition, automated tattoo recognition, and ALPRs could be limited if private companies are allowed to claim that the technology is protected by an expanded Exemption 4.
“The public has a strong interest in accessing information related to which types of facial recognition software the government is using, in what ways it is being used, and the level of accuracy achieved,” the brief argues. “Expanding Exemption 4 could frustrate these goals by allowing companies to self-designate records as ‘confidential.’”
The brief also demonstrates how it is essential that FOIA permit access to information about how the government makes important decisions about benefits it provides to the public or other uses of automated decision-making skills, as those decisions increasingly rely on private technology, including algorithms.
“These algorithms purport to make the state’s allocation of scarce resources more efficient, but they are easily infected with grave defects,” the brief states. “In one case, a court found that the state’s automated Medicaid budgeting system was so unreliable that it ‘arbitrarily deprive[d] participants of their property rights and hence violate[d] due process.’”
Joining EFF on the brief were New York University’s AI Now Institute, the American Civil Liberties Union, the Center for Race, Inequality, and the Law at New York University School of Law, and the Knight First Amendment Institute at Columbia University.
EFF would like to thank the University of Buffalo School of Law’s Civil Liberty and Transparency Clinic for writing and filing the brief. We owe a particular debt of gratitude to student attorneys Suzanne Starr, John Zakour, and John Kueble, as well as Assistant Clinical Professor Jonathan Manes, who supervised their work.