LSC's 2020 Innovations in Technology Conference

Algorithmic decision-making has enormous potential to help vulnerable people in our society. By reducing the role of human decision-makers, Artificial Intelligence (AI) promises to increase efficiencies, reduce biases, and simplify processes. States want AI to decide eligibility for Medicaid, public housing, SNAP benefits, and more. But, with all this promise and potential, why are we seeing litigation challenging automated decision-making in Arkansas, Colorado, Michigan, New York and other states? The lawsuits are contesting applications that have gone awry with erroneous outcomes, embedded biases, and wrongful findings of fraud, and are insisting, among other things, that government adjust processes and standards, and re-do adverse decisions. AI applications are going awry with erroneous decisions, embedded biases, wrongful findings of fraud, and requirements to re-do decisions. Will AI systems, including portals and online dispute resolution, work as hoped? Due process, equal protection, Freedom of Information Act (FOIA), and statutory challenges are proceeding against AI models, and are on the rise. What are the consequences of saying to ourselves and to others “the computer did it”? Our community needs advocacy tools, including national standards, to hold government, tech firms, and other powerful interests, accountable to the rule of law. We must preserve the rights of people against the new tyranny of presumptions of accuracy. Join this group conversation about the promise, the perils, and the national standards that are defining the entry of AI into the justice sphere.