By Noah Bunnell
Algorithms increasingly play a central role in the provision of public benefits, offering government entities previously unimaginable ways of optimizing public services, but they also pose risks of error, bias, and opacity in government decision-making. At present, many publicly-deployed algorithms are created by private companies and sold to government agencies. Given robust protections for trade secrets in the courts and feeble state open records laws, such algorithms, even those with fundamental flaws or biases, may escape regulatory scrutiny. If state and local governments are to avail themselves of the benefits of algorithmic governance without triggering its potential harms, they will need to act quickly to design regulatory systems that are flexible enough to respond to continual innovation yet durable enough to withstand regulatory capture. This Note proposes a novel regulatory solution in the form of a new, independent agency at the state or local level — an Algorithmic Transparency Commission — devoted to the regulation of publicly-deployed algorithms. By establishing such an agency, tailored to the needs of each jurisdiction, state and local governments can continue to enhance their efficiency and safeguard companies’ proprietary information, while also fostering a greater degree of algorithmic transparency, accountability, and fairness.