Bridging the Transparency Gap: What Can Explainable AI Learn from the AI Act?

The European Union has proposed the Artificial Intelligence Act which introduces detailed requirements of transparency for AI systems. Many of these requirements can be addressed by the field of explainable AI (XAI), however, there is a fundamental difference between XAI and the Act regarding what transparency is. The Act views transparency as a means that supports wider values, such as accountability, human rights, and sustainable innovation. In contrast, XAI views transparency narrowly as an end in itself, focusing on explaining complex algorithmic properties without considering the socio-technical context. We call this difference the ``transparency gap''. Failing to address the transparency gap, XAI risks leaving a range of transparency issues unaddressed. To begin to bridge this gap, we overview and clarify the terminology of how XAI and European regulation -- the Act and the related General Data Protection Regulation (GDPR) -- view basic definitions of transparency. By comparing the disparate views of XAI and regulation, we arrive at four axes where practical work could bridge the transparency gap: defining the scope of transparency, clarifying the legal status of XAI, addressing issues with conformity assessment, and building explainability for datasets.

[1]  Agathe Balayn,et al.  Explainability in AI Policies: A Critical Review of Communications, Reports, Regulations, and Standards in the EU, US, and UK , 2023, FAccT.

[2]  J. Ser,et al.  Explainable Artificial Intelligence (XAI): What we know and what is left to attain Trustworthy Artificial Intelligence , 2023, Inf. Fusion.

[3]  Tim Miller.  Explainable AI is Dead, Long Live Explainable AI!: Hypothesis-driven Decision Support using Evaluative AI , 2023, FAccT.

[4]  Shay B. Cohen,et al.  Causal Explanations for Sequential Decision-Making in Multi-Agent Systems , 2023, AAMAS.

[5]  Sandra Wachter,et al.  Trustworthy artificial intelligence and the European Union AI act: On the conflation of trustworthiness and acceptability of risk , 2023, Regulation & Governance.

[6]  Esther Keymolen,et al.  Explanation and Agency: exploring the normative-epistemic landscape of the “Right to Explanation” , 2022, Ethics and Information Technology.

[7]  T. Seidel,et al.  Towards Human-centered Explainable AI: User Studies for Model Explanations , 2022, ArXiv.

[8]  Finale Doshi-Velez,et al.  Connecting Algorithmic Research and Usage Contexts: A Perspective of Contextualized Evaluation for Explainable AI , 2022, HCOMP.

[9]  S. Saralajew,et al.  A Human-Centric Assessment Framework for AI , 2022, ArXiv.

[10]  Jakob Schoeffer,et al.  “There Is Not Enough Information”: On the Effects of Explanations on Perceptions of Informational Fairness and Trustworthiness in Automated Decision-Making , 2022, FAccT.

[11]  M. Palmirani,et al.  Metrics, Explainability and the European AI Act Proposal , 2022, J.

[12]  Kaley J. Rittichier,et al.  Trustworthy Artificial Intelligence: A Review , 2022, ACM Comput. Surv..

[13]  L. Chen,et al.  CPKD: Concepts-Prober-Guided Knowledge Distillation for Fine-Grained CNN Explanation , 2021, 2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT).

[14]  M. Cannarsa.  Ethics Guidelines for Trustworthy AI , 2021, The Cambridge Handbook of Lawyering in the Digital Age.

[15]  N. Jennings,et al.  Trustworthy human-AI partnerships , 2021, iScience.

[16]  Plamen P. Angelov,et al.  Explainable artificial intelligence: an analytical review , 2021, WIREs Data Mining Knowl. Discov..

[17]  Michael Veale,et al.  Demystifying the Draft EU Artificial Intelligence Act , 2021, ArXiv.

[18]  Bodhisattwa Prasad Majumder,et al.  Knowledge-Grounded Self-Rationalization via Extractive and Natural Language Explanations , 2021, ICML.

[19]  Gesina Schwalbe,et al.  A comprehensive taxonomy for explainable artificial intelligence: a systematic survey of surveys on methods and concepts , 2021, Data Mining and Knowledge Discovery.

[20]  Marcin Detyniecki,et al.  Understanding Prediction Discrepancies in Machine Learning Classifiers , 2021, ArXiv.

[21]  C. Rudin,et al.  Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges , 2021, Statistics Surveys.

[22]  Emre Bayamlıoğlu.  The right to contest automated decisions under the General Data Protection Regulation : Beyond the so‐called “right to explanation” , 2021 .

[23]  Michael Winikoff,et al.  Artificial Intelligence and the Right to Explanation as a Human Right , 2021, IEEE Internet Computing.

[24]  Holger Hermanns,et al.  What Do We Want From Explainable Artificial Intelligence (XAI)? - A Stakeholder Perspective on XAI and a Conceptual Model Guiding Interdisciplinary XAI Research , 2021, Artif. Intell..

[25]  Marco F. Huber,et al.  A Survey on the Explainability of Supervised Machine Learning , 2020, J. Artif. Intell. Res..

[26]  Mireille Hildebrandt,et al.  Law for Computer Scientists and Other Folk , 2020 .

[27]  Vlad I. Morariu,et al.  Black-box Explanation of Object Detectors via Saliency Maps , 2020, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[28]  Aaron Sedley,et al.  Exciting, Useful, Worrying, Futuristic: Public Perception of Artificial Intelligence in 8 Countries , 2019, AIES.

[29]  Peter A. Flach,et al.  Explainability fact sheets: a framework for systematic assessment of explainable approaches , 2019, FAT*.

[30]  Brandon M. Greenwell,et al.  Interpretable Machine Learning , 2019, Hands-On Machine Learning with R.

[31]  Jaime S. Cardoso,et al.  Machine Learning Interpretability: A Survey on Methods and Metrics , 2019, Electronics.

[32]  Gary Klein,et al.  Metrics for Explainable AI: Challenges and Prospects , 2018, ArXiv.

[33]  Thomas Lukasiewicz,et al.  e-SNLI: Natural Language Inference with Natural Language Explanations , 2018, NeurIPS.

[34]  Eric D. Ragan,et al.  A Multidisciplinary Survey and Framework for Design and Evaluation of Explainable AI Systems , 2018, ACM Trans. Interact. Intell. Syst..

[35]  Chris Russell,et al.  Explaining Explanations in AI , 2018, FAT.

[36]  M. Kaminski.  The right to explanation, explained , 2018, Research Handbook on Information Law and Governance.

[37]  Timnit Gebru,et al.  Datasheets for datasets , 2018, Commun. ACM.

[38]  Michael Veale,et al.  Enslaving the Algorithm: From a “Right to an Explanation” to a “Right to Better Decisions”? , 2018, IEEE Security & Privacy.

[39]  Roland Vogl,et al.  Rethinking Explainable Machines: The GDPR's 'Right to Explanation' Debate and the Rise of Algorithmic Audits in Enterprise , 2018 .

[40]  Franco Turini,et al.  A Survey of Methods for Explaining Black Box Models , 2018, ACM Comput. Surv..

[41]  Deborah G. Johnson,et al.  Reframing AI Discourse , 2017, Minds and Machines.

[42]  Giovanni Comandé,et al.  Why a Right to Legibility of Automated Decision-Making Exists in the General Data Protection Regulation , 2017 .

[43]  Tim Miller,et al.  Explanation in Artificial Intelligence: Insights from the Social Sciences , 2017, Artif. Intell..

[44]  Luciano Floridi,et al.  Why a Right to Explanation of Automated Decision-Making Does Not Exist in the General Data Protection Regulation , 2017 .

[45]  Tiffany Curtiss.  [91WashLRev1813] Computer Fraud and Abuse Act Enforcement: Cruel, Unusual, and Due for Reform , 2016 .

[46]  Seth Flaxman,et al.  European Union Regulations on Algorithmic Decision-Making and a "Right to Explanation" , 2016, AI Mag..

[47]  Marco Tulio Ribeiro,et al.  "Why Should I Trust You?": Explaining the Predictions of Any Classifier , 2016, HLT-NAACL Demos.

[48]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[49]  B. Malle,et al.  How People Explain Behavior: A New Theoretical Framework , 1999, Personality and social psychology review : an official journal of the Society for Personality and Social Psychology, Inc.

[50]  R. Schifter.  White House , 1996 .

[51]  Andrew Sheppard,et al.  Parliament , 1982, The Lancet.

[52]  L. Agustín,et al.  European Parliament , 1979, International and Comparative Law Quarterly.

[53]  G. Williams.  Causation in the Law , 1961, The Cambridge Law Journal.

[54]  Sandra Wachter,et al.  Trustworthy Artificial Intelligence and the European Union AI Act: On the Conflation of Trustworthiness and the Acceptability of Risk , 2022, SSRN Electronic Journal.

[55]  Open Rights Group response to the DCMS policy paper “Establishing a pro-innovation approach to regulating AI” , 2022 .

[56]  Ilia Stepin,et al.  A Survey of Contrastive and Counterfactual Explanation Generation Methods for Explainable Artificial Intelligence , 2021, IEEE Access.

[57]  P. Hacker,et al.  Varieties of AI Explanations Under the Law. From the GDPR to the AIA, and Beyond , 2020, xxAI@ICML.

[58]  Maël Pégny,et al.  The Right to an Explanation , 2019, Delphi - Interdisciplinary Review of Emerging Technologies.

[59]  Marco Tulio Ribeiro,et al.  “ Why Should I Trust You ? ” Explaining the Predictions of Any Classifier , 2016 .

[60]  Dear Mr Sotiropoulos.  ARTICLE 29 Data Protection Working Party , 2013 .

[61]  Mireille Hildebrandt,et al.  The Dawn of a Critical Transparency Right for the Profiling Era , 2012 .

[62]  Gunther Teubner.  Breaking Frames: The Global Interplay of Legal and Social Systems , 1997 .

[63]  John Mingers,et al.  Law as an Autopoietic System , 1995 .

[64]  Hengshuai Yao,et al.  Explainable Artificial Intelligence for Autonomous Driving: A Comprehensive Overview and Field Guide for Future Research Directions , 2021, ArXiv.