Generative AI, Explainability, and Score-Based Natural Language Processing in Benefits Administration

Authors

  • Frank Pasquale Cornell University
  • Gianclaudio Malgieri University of Leiden

Keywords:

generative AI, AI, GenAI, natural language processing, automation, benefits, administration, automated decisionmaking, rule of law, administrative law, disability, social security, public assistance, machine learning, ML, ADM

Abstract

Administrative agencies have developed computationally-assisted processes to speed benefits to persons with particularly urgent and obvious claims. One proposed extension of these programs would score claims based on the words that appear in them, identifying some set of claims as particularly like known, meritorious claims, without understanding the meaning of any of these legal texts. Score-based natural language processing (SBNLP) may expand the range of claims that may be categorized as urgent and obvious, but as its complexity advances, its practitioners may not be able to offer a narratively intelligible rationale for how or why it does so. At that point, practitioners may utilize the new textual affordances of generative AI to attempt to fill this explanatory gap, offering a rationale for decision that is a plausible imitation of past, humanly-written explanations of judgments.

This article explains why such generative AI should not be used to justify SBNLP decisions in this way. Due process and other core principles of administrative justice require humanly intelligible identification of the grounds for adverse action. Given that ‘next-token-prediction’ is distinct from understanding a text, generative AI cannot perform such identification reliably. Moreover, given current opacity and potential bias in leading chatbots based on large language models, as well as deep ethical concerns raised by the databases they are built on, there is a good case for entirely excluding these automated outputs in administrative and judicial decision-making settings. Nevertheless, SBNLP may be established parallel to or external to justification-based legal proceedings, for humanitarian purposes.

Author Biography

Gianclaudio Malgieri, University of Leiden

Dr. Gianclaudio Malgieri is an Associate Professor of Law & Technology and a Board Member at eLaw - Center for Law and Digital Technologies. He serves as the Co-Director of the Brussels Privacy Hub, Free University of Brussels (VUB) and as an Affiliated Researcher at the Augmented Law Institute of the EDHEC Business School (Lille, France). He is an Associate Editor of Computer Law and Security Review, an External Ethics Expert of the European Commission, and an Advisory Board member of EPIC.org. He also coordinates “VULNERA“, the International Observatory of Vulnerable People in Data Protection. He conducts research on and teaches Data Protection Law, privacy, AI regulation, Digital Law, Consumer protection in the digital market, Data Sustainability, Intellectual Property Law.

Gianclaudio has authored more than 60 publications, including articles in leading international academic journals.  His works have been cited by, inter alia, top international newspapers (The New York TimesThe Washington PostLe MondePoliticoLa TribuneFrance CultureilSole24Ore, la Repubblica, il Corriere della Sera, Euractiv) ) but also institutions, e.g. the European Commission and the Council of Europe, the World Economic Forum, the Canadian Government, and the Canadian Data Protection Authority.  In 2020 he was the only EU scholar to receive the FPF Privacy for Policymaker Paper Award. He published in English, Italian and French and some works were translated even into Chinese.

He got an LLM with honours at the University of Pisa (2016) and a JD with honours at S.Anna School of Advanced Studies of Pisa (2017). He was visiting student at the London School of Economics (2013), the World Trade Institute of the University of Bern (2014), École Normale Superieure de Paris (2015) and Oxford University (2018).  He obtained a PhD in Law at the Law, Science, Technology and Society (LSTS) Research Centre of the Vrije Universiteit Brussel, where he is now a Guest Professor and Affiliated Researcher. His PhD thesis focused on the notion of data subjects in the GDPR, in particular on the vulnerable data subjects. He also conducts research on automated decision-making, privacy and fundamental rights, surveillance, data ownership, intellectual privacy, and consumer law.

Downloads

Published

20 May 2024
Total downloads
150

How to Cite

Pasquale, Frank, and Gianclaudio Malgieri. 2024. “Generative AI, Explainability, and Score-Based Natural Language Processing in Benefits Administration”. Journal of Cross-Disciplinary Research in Computational Law 2 (2). https://journalcrcl.org/crcl/article/view/59.