IPIE Announces Scientific Panel on Global Standards for AI Audits
January 2024
Developing global standards for auditing AI systems to build responsible AI and ensure equitable access to information technology.
The International Panel on the Information Environment (IPIE) has established a new Scientific Panel on Global Standards for AI Auditing which will build guidance on how public-facing AI systems should be inspected and identify the protocols that should be used.
Drawing on the latest knowledge from the social, computer, and engineering sciences—as well as the humanities—the purpose of this scientific panel is to build consensus about the operationalization of AI auditing standards, about what is desirable and practicable in auditing public-facing AI systems.
Leading scientists and policy experts from the AI field will form the panel, drawn from different (sometimes conflicting) disciplines and geographies. It will be chaired by Professor Wendy Hui Kyong Chun, a trailblazing new media theorist and Director of the Digital Democracies institute at Simon Fraser University.
Professor Wendy Hui Kyong Chun said: “There has been snowballing interest in the possibilities and challenges of AI to the information environment and its regulation. Audits are the key to enforcing regulation.
“It goes without saying that industry, government, and civil society have very different expectations for what constitutes an AI audit, when it should be implemented, and what consequences should follow from the outcomes of an audit.
“To advance this global conversation, the IPIE will help define the critical, comprehensive, and reasonable criteria for AI auditing—standards that themselves can be globally applied. “
Panel member and Professor Alondra Nelson, former acting director of the White House Office of Science and Technology Policy said:
“Evidence-based, consensus-driven guidance on AI governance is urgently needed and this Panel will be at the forefront of that guidance. In this major election year, disinformation and misinformation are going to be significant problems, and we need the right tools to meet the additional challenges posed by the growing uses of AI tools and systems. AI holds the potential to redefine every part of our society and potentially benefit society, but only if we get the regulation right. Global standards on AI audits have a major role to play in AI governance.”
Fellow panel member Gbenga Sesan, an expert in digital inclusion and rights across Africa said: “The panel's work on global standards for AI auditing will help to ensure AI systems are transparent, fair, and accountable. The work will take inclusive approaches to AI auditing that involve diverse stakeholders, including marginalized communities, to address potential biases and ensure that AI technologies benefit all segments of society.”
Why a Focus on AI Auditing?
While we may not fully understand the AI systems that shape our lives, many of us have become dependent on their use. Unfortunately, the number of poor design choices, clear cases of discrimination, lost opportunities for constructive choices, and unexplained outcomes has grown. Several regulators have identified public-facing AI systems that have a significant impact on the life course, and life opportunities, available to citizens and consumers.
Auditing an algorithm can range from reviewing the code of a machine learning system, to understanding the source and application of training data, to examining the outputs and social impacts of AI systems.
There are dozens of ethical guidelines, normative frameworks, and value statements about how AI should be designed and implemented. At the same time, the firms that design such tools consistently provide little or no information about the practical decisions taken, the provenance of training datasets, or details beyond a pledge to build publicly facing machine learning systems aligned with the social good.
There remains no global criteria guiding standards for auditing AI systems, despite calls from industry and government to produce them. There are countless private sector auditing services with uncertain standards, and a growing expectation within several legal jurisdictions that such auditing systems be available and adhered to.
The Scientific Panel on Global Standards for AI Audits seeks to achieve an expansive, global, and cross disciplinary perspective on AI auditing through expert consensus.
What will the Panel Do?
The purpose of this scientific panel is to build consensus about the operationalization of AI auditing standards, using the latest knowledge from the humanities and social, computer, and engineering sciences about what is desirable and practicable in auditing public-facing algorithms.
The overall objectives of the IPIE require all outputs to explicitly advance the implementation of algorithmic audits for:
- understanding how researcher access to data provides what is needed for algorithmic audits;
- what measurements and indices are required to improve AI auditing standards, so they both engage global impacts and local contexts;
- assessing the guiding questions and aims of AI auditing at a global scale.
There are several national auditing standards that may be suitable for elevation to global standards, such as those developed in the UK, as well as those developed by the ISO. Building a trustworthy AI assurance ecosystem will require grounding and validation in science and engineering.
An important part of this work will involve evaluating the proposals from civil society, industry, and government—along with those from scientific and engineering communities—to establish the expectations for data and code provenance. Establishing standards for data provenance is key in AI auditing, and highly relevant for sourcing synthetic media, mis/disinformation, deep fakes, and now, generative AI.
Outputs
Outputs from this Scientific Panel will be strategically timed to feed into the relevant regulatory frameworks that are in need of authoritative testimony, amicus briefs, and other modes of expert engagement.
The panel will produce technical papers with timely policy briefs with the following themes:
- A literature review and gap analysis on the landscape of AI audits;
- A paper on the prospects for data provenance and researcher access to data;
- Criteria, evaluation, evidentiary documentation, and processes crucial to establishing global standards for AI auditing.
The papers will be drafted by our consulting scientists, guided by our panel members, and will be posted on the Reports page and the Panel's page of the IPIE website.
Who is on the Scientific Panel on Global Standards for AI Audits?
Chair: Prof. Wendy Hui Kyong Chun (Canada) - Simon Fraser University’s Canada 150 Research Chair in New Media, Professor in the School of Communication, Director of the Digital Democracies Institute, and Fellow of the Royal Society of Canada. She has served as a Commissioner for the Canadian Commission on Democratic Expression, a member of the Canadian Council of Canadian Academies Expert Panel on AI for Science and Engineering, and the Royal Society of Canada Working Group on Protecting Public Advice. She is author of eight books, and has held fellowships from numerous foundations and Institutes, including the Guggenheim and the Institute for Advanced Study (Princeton).
Vice Chair: Prof. Anikó Hannák (Hungary), Assistant Professor at the computer science department of the University of Zürich. Broadly, her work investigates a variety of content serving websites such as Search Engines, Online Stores, Job Search Sites or Freelance Marketplaces. In her PhD work she created a methodology called Algorithmic Auditing which tries to uncover the potential negative impacts of large online systems.
Vice Chair: Prof. Marcelo Mendoza (Chile), Associate Professor of the Department of Computer Science, Pontifical Catholic University of Chile. He is an Electronic Engineer and Master of Science in Informatics of the Technical University Federico Santa María (2000) and Doctor in Sciences with Mention in Computation of the University of Chile (2007). He worked as a postdoctoral researcher in Yahoo!Research. He was president of the Chilean Association of Pattern Recognition (ACHIRP) from 2015 to 2018. His areas of interest in research are text mining, information recovery and data mining in social networks.
Prof. Meredith Broussard (United States), Associate Professor at the Arthur L. Carter Journalism Institute of New York University, research director at the NYU Alliance for Public Interest Technology, and the author of several books, including “More Than a Glitch: Confronting Race, Gender, and Ability Bias in Tech” and “Artificial Unintelligence: How Computers Misunderstand the World.” Her academic research focuses on artificial intelligence in investigative reporting and ethical AI, with a particular interest in using data analysis for social good.
Prof. Ruben Enikolopov (Russia), ICREA Research Professor at UPF and Barcelona Institute of Political Economy and Governance and a visiting professor at the New Economic School. His research interests include political economy, economics of mass media, and development economics. His work is mainly focused on empirical analysis related to political economy, economics of mass media, development economics, and corporate finance.He has published in American Economic Review, Quarterly Journal of Economics, Review of Economic Studies, Econometrica, Proceedings of National Academy of Science, American Political Science Review, American Economic Journal: Applied Economics. He holds a PhD in Economics from Harvard University, MA in Economics from New Economic School and MSc in Physics from Moscow State University.
Dr. Alondra Nelson (United States), Acclaimed scholar and writer Alondra Nelson is the Harold F. Linder Professor at the Institute for Advanced Study. Dr. Nelson served as deputy assistant to President Joe Biden and acting director of the White House Office of Science and Technology Policy (OSTP). As a distinguished senior fellow at the Center for American Progress, and a science and technology policy advisor, she has provided guidance to local, state, and federal governments, multilateral and intergovernmental organizations, legislators, civil society, and others. Including Nelson in the list of Ten People Who Shaped Science, Nature said of her OSTP tenure, “this social scientist made strides for equity, integrity and open access.” In 2023, she was included in the inaugural TIME100 list of the most influential people in artificial intelligence. She is the author of several books, most recently The Social Life of DNA, an award-winning exploration of the social implications of direct-to-consumer genetic technologies.
Prof. Christian Sandvig (United States), Visiting Professor in Social Sciences at the Institute for Advanced Study and H. Marshall McLuhan Collegiate Professor of Digital Media appointed in the Department of Communication & Media and the School of Information at the University of Michigan. He is also the Director of ESC: The Center for Ethics, Society, and Computing. He is a social researcher studying the consequences of algorithmic systems that curate and organize culture. He is recently known for research on "algorithm auditing."
Mr. Gbenga Sesan (Nigeria), Executive Director of Paradigm Initiative, a pan-African social enterprise focusing on digital inclusion and rights. He is also a Non-Resident Fellow at Stanford University's Digital Civil Society Lab. With extensive consulting experience for organizations like Microsoft and Harvard University in over 30 countries, he’s earned recognition as a Schwab Foundation Social Entrepreneur of the Year and held various fellowships. Mr Sesan has advised UN committees and governmental bodies on technology and broadband infrastructure. His expertise extends to data privacy, as seen in their advisory roles for organizations like the World Health Organization and the World Economic Forum. Additionally, he’s been recognized by CNN and Ventures Africa for their contributions to technology and leadership in Africa.
Prof. Weixing Shen (China) - Professor in the field law and technology at Tsinghua University. He was Fulbright Visiting Scholar at Harvard Law School and the Humboldt Scholar at Freiburd University.
Mr. Andrew Sporle (New Zealand), Senior Research Fellow at the University of Auckland's Department of Statistics and the Managing Director at iNZight Analytics Ltd. With extensive experience spanning over two decades, he specializes in social and health research and Māori research workforce development. His recent focus involves official statistics research across various sectors, including public, private, and academic domains. Andrew's statistical expertise is rooted in epidemiology, with advanced training in genetic and epigenetic epidemiology.
Prof. Rohini Srihari (United States), Professor in the Dept. of Computer Science and Engineering at the State University of New York, Buffalo where her research focuses on natural language processing, artificial intelligence, machine learning and information retrieval. Her recent research focuses on advancing the state-of-the-art in socialbots capable of engaging in empathetic, interesting and purposeful conversations. She directs a research group focused on Conversational AI for Social Good that addresses research issues in building trustworthy socialbots for various purposes, including combating disinformation as well as assisting the disabled. She has also worked extensively with the US Government in developing innovative multilingual text mining solutions.
Prof. Christo Wilson (United States), Associate Professor and Associate Dean of Undergraduate Programs at Northeastern University's Khoury College of Computer Sciences. He is a founding member of the Cybersecurity and Privacy Institute there. With a PhD from the University of California, Santa Barbara, under Professor Ben Y. Zhao, his research focuses on the intersection of Big Data, security, and privacy, drawing from various scientific disciplines. Wilson has received prestigious awards and fellowships, including being a 2019 Sloan Fellow and a 2019-2020 Fellow at the Berkman Klein Center for Internet & Society. His work is supported by grants from institutions such as the NSF, Sloan Foundation, and Mozilla Foundation. Notably, he has received best paper awards at SIGCOMM, NDSS, and ICWSM conferences and has been featured in prominent media outlets like CBS Evening News and The Wall Street Journal. Wilson actively contributes to academic communities.
Prof. Mimi Zou (China/United Kingdom), holds the chair in commercial law and serves as Senior Academic Lead and Co-Director of the Centre for Commercial and Private Law at Exeter Law School. Her expertise lies in comparative contract, commercial, and employment law, with a focus on emerging technologies such as AI, blockchain, and digital assets. Previously at the University of Oxford, she established and led the lawtech innovation lab. Her research has been published in top journals and earned international recognition, including prestigious prizes, awards, and fellowships. As a dual-qualified lawyer with extensive international experience, she has held advisory roles at notable organizations like the UK Government's Centre for Data Ethics and Innovation, UNIDROIT Working Group on Private Law and Digital Assets, and the World Economic Forum Experts Network.