Digital Identity During Times of Crisis
What We Learned During the Fall 2022 Research Sprint
The Berkman Klein Center for Internet & Society hosted a 10-week Research Sprint from October to December 2022 investigating Digital Identity in Times of Crisis, in collaboration with partners metaLAB at Harvard, the Edgelands Institute, and Access Now. BKC Research Sprints are an educational format developed at the Center that connects early-career scholars and practitioners with leading subject matter experts and stakeholders to troubleshoot specific social, ethical, and policy challenges related to digital technology. Each Sprint culminates in the production of outputs developed with the guidance of external experts and a partner organization.
For the purposes of this Sprint, digital identity refers to the methods, systems, and policies used to verify and authenticate people over digital channels, as well as sociotechnical systems used to remotely identify individuals. This broad perspective includes digital IDs issued by both governmental and private institutions and a wide range of technologies, such as the universally familiar username-password system, biometrics, and decentralized identity solutions. The adoption of digital ID systems to mediate access to fundamental aspects of public and private life holds great promise as well as risks. Understanding and scrutinizing the impacts of specific policy and design choices, particularly on historically vulnerable or crisis-impacted communities, was a core objective of the Sprint.
Members of the global research sprint cohort included 33 early career young scholars and professionals based in 17 countries across five continents. Their backgrounds included experience in law, public policy, social work, philosophy, and other disciplines.
The Research Sprint on Digital Identity was ambitious, involving a greater number of partner organizations than previous sprints and dividing participants into three distinct methodological approaches. Led by Dr. Kim Albrecht, one group worked with designers and data scientists on data visualization projects. The second group, led by Dr. Amy Johnson, a linguistic anthropologist, created works of speculative fiction. The last cohort worked in small groups alongside advisors Santiago Uribe Saenz and Laura García Vargas from the Edgelands Institute, Jad Esber a BKC Affiliate, and Adam Nagy a BKC staff member to produce policy documents or web-based resources.
Why Digital ID?
The world is confronting overlapping crises: a pandemic, climate-driven disasters, international conflicts, democratic backsliding, and global inequality. Under such circumstances, decision-makers are under pressure to make impactful decisions quickly. Quick digital identity solutions can backfire, deepening societal inequities, jeopardizing data security, and eroding privacy through new avenues of surveillance. For instance, the Aadhaar identification system in India, one of the largest biometric digital identity systems, has had frequent security breaches, and has been criticized by human rights groups for providing the government with a new tool for surveillance and social control. Vulnerabilities are a persistent reality for any technical enterprise, but with appropriate transparency policies, contigency planning, and resources for remediating harms, such risks can be managed. In Estonia, a country that has had a digital ID system in place since 2001, there have been major security breaches over the years. In each instance, the government has made issues public, acted decisively to resolve security threats and pursued accountability when appropriate. On balance, Estonian citizens have benefited from the decreased administrative costs and improved access to services afforded to them by their digital identification system.
The research sprint began with opening sessions defining digital identity and mapping key stakeholders before focusing on understanding potential risks and benefits of digital identity systems through discussions with technical and policy experts. The program culminated with a final event reimagining digital identity for the future. Throughout the sprint, participants worked collaboratively to produce public-facing outputs. The production of these outputs was divided into three methodological categories: Data Visualization, Speculative Fiction, and Public Policy. Each product is intended to inform an identified stakeholder or audience about a policy issue related to digital identity, but the different methodological tracks enabled participants to use vastly different skills and approaches to relay their message.
Data Visualization Track
Internet users leave a steady trail of digital footprints in their wake. Research sprint participants in the data visualization cohort interrogated the portrayals constructed by these data traces, exploring how data collected by search engines, social media sites, and other platforms can amalgamate into pseudo identities, used for a variety of commercial ends, and what such processes neglect.
Under the guidance of Dr. Kim Albrecht, as well as assistance from Justin Clark, Dario Rodighiero, and Jonas Parnow, participants identified services that create digital identities to track individuals, gathered data (one group used public datasets while another requested data about themselves), and created visual representations to interrogate the portrayals created by digital data. The group that used data about themselves compared the accuracy of inferential user-interest profiles created by social media companies against judgments about their interests from their own friends and family. The other group used population-level data in conjuction with individual stories of people struggling to use digitial identity systems as a method to highlight the need to ground our perception of broad trends in an understanding of the “outlier” cases and to inform policy-conversations with real-world examples. Through these different approaches both groups challenge us to reflect on what data reveals and what it might obscure or exclude.
The “Digital Banking Team” produced a report titled “Lost in Digitalization: What Data Reveals and What it Hides – Stories of Digital Banking and Digital Identity From India.” The report presents vignettes of digital identity-based challenges faced by five fictionalized users of financial digital services in India. For example, in one vignette, a worker is unable to login into a biometric identification system because through aging and hard labor his fingerprints are not recognized. In a different example, the authors describe the struggle Raghu, a fictionalized cart-puller in Hyderabad, faces using a digital identity-based system for payments from customers when his access to Internet data is limited and is a significant financial burden. Throughout the report each vignette is bolstered by data visualizations that contexualize the issue explored in the individual story. For instance Raghu’s story is accompanied by a line graph showing information about Internet access in India over time and a block chart estimating Raghu’s monthly expenses.
The other team focused on social media profiling. In “Reality vs. Social Media: An Exploration of Profiling,” sprint participants requested information about themselves from popular social media platforms such as Twitter, Facebook, and LinkedIn, which all make inferences about user interests. The participants surveyed a group of their friends, family, and colleagues to determine whether they thought the inferred interests from the social media platforms were accurate. Each team member created a visualization depicting how well the platforms’ inferred interests aligned with the judgment of the survey respondents and how well the survey results reflected their true preferences. They found that survey participants were able to discern the legitimacy of the inferred interests but the inferred interests identified by the platforms were inaccurate.
Four groups of policy track participants created resources for digital identity stakeholders. Each group determined their particular topic, audience, and goals.
The Digital ID Threat Modeling group, with light guidance from BKC affiliate Jad Esber, created a website and online tool to assist stakeholders in evaluating possible threats to the issuance and implementation of a national, government-issued digital identity system. The threat modeling tool asks users to select among several personas (an advocate engaged in expanding access to social services, a direct service provider of social services, and a governmental policy-maker) followed by either a concern or their role in the lifecycle of digital identity system (e.g., designer, registrar, user, manager etc.). Using this information, the tool presents short descriptions of the potential challenges and links to relevant external resources.
Guided by Santiago Uribe of Edgelands Institute, the second group documented and synthesized principles, guidelines, and good practices for inclusivity and accessibility in digital identification systems. Their report, titled “Enhancing Inclusion in Digital Identity Policies and Systems: An Assessment Framework,” was the result of a content analysis of a corpus of 42 digital identification policy documents that referenced principles of inclusivity and accessibility. Through this mapping and analytical process, the team derived key aspects and commonalities across the policy landscape which they synthesized into “7 Good Practices to Foster Inclusion.”
In the context of digital identification systems, discretion broadly refers to policies, practices, and design choices that enable the use of human judgment. Guided by Laura Garcia from the Edgelands Institute, research sprint participants in policy group three created a primer for policymakers and industry professionals on the fundamental role discretion (e.g., human oversight systems, user autonomy and choice, flexibility to override decisions) plays in the design and management of robust, inclusive, and flexible digital identification systems. Using real-world examples, the booklet defines the importance of discretion, discusses benefits and risks associated with different degrees and types of discretion, and explores discretion within the context of regulatory frameworks in the European Union, the United Kingdom, and Kenya. The booklet also delves into the interaction of discretion with different digital identification technologies and culminates in recommendations to policy makers and digital identity system developers.
The final group, with guidance from BKC staff member Adam Nagy, created an interactive risk assessment tool intended to help specific vulnerable populations (and relevant civil society organizations) understand possible risks associated with digital ID systems, learn about possible ameliorative policy solutions, and connect with additional resources and external organizations working on related issues. The chat-based tool varies its responses depending on information provided by the user, specifically the user’s identification with one of four possible categories: a woman living in a rural area, membership in a gender or sexual minority, presence in a One Belt, One Road Initiative country, or being a child.
Speculative Fiction Track
By presenting ideas that do not exist in the real world, speculative fiction allows us to imagine new possibilities. It prompts us to take a step back from the realities in which we’re embedded to negotiate new ones, and in doing so, it offers us a fresh vantage point from which to consider our own world. As output track lead Dr. Amy Johnson wrote in her introduction to the speculative fiction anthology sprint participants created, “Storytelling is at once both a simple, familiar tool and a complex one, for when we explore through story, we must work with character and context and change. Writing and reading speculative fiction thus surface fresh insights and questions—and, perhaps most importantly, give us a means not only to know but to understand.”
Under the guidance of Dr. Johnson, with assistance from mentors Madeline McGee and Dr. Lis Sylvan, participants in the speculative fiction track worked both individually and as a cohort to create a collection of short fiction stories exploring digital identity. Stories from (Un)Identified Worlds: A Speculative Anthology examines, sometimes directly and sometimes through metaphor, the lives that shape and are shaped by digital identity systems and the power hierarchies within which these systems interact. The authors were challenged to approach digital identity from a “hopepunk” perspective. Neither utopian nor dystopian, hopepunk rejects both cynicism and naivety and embraces the process of perseverance and resistance, which often lie in community cooperation and radical kindness. In a hopepunk story, the world is never too broken to change, but its problems may not be cleanly resolved in the end.
The anthologized stories, which feature sky whales, mist-based artificial general intelligences, free-thinking databots, martians, and other eclectic characters, speak to the complex layering of digital identity systems and the multifaceted ways these systems can not only exclude and target people, but also offer them new ways asserting their dignity and affirming their humanity. The lives of the characters in these stories bring a wide range of perspectives to their interactions with identity systems and the characters are instantly recognizable as very much like ourselves–ordinary people who must grapple with inflexible systems that cannot recognize or adapt to the multiplicity of human identity, people who take advantage of these systems for their own ends, people who fight against control and exclusion, and people who come together to construct new systems based on trust and communal well being. Together, the characters and the stories give us new ways to think about the digital identity systems that exist in our world and an avenue for imagining better and brighter futures ahead.
For many participants, their work on digital identity continues. Mardiya Siba Yahaya, Gabriel Fonlladosa, Lilian Olivia Orero, and Sean Chen, will present their interactive online risk assessment tool at MozFest in March 2023. Mariana Rozo-Paz published three blog posts for the Datasphere Initiative drawing from her experience as a sprint participant.
For a complete list of sprint participants and to follow their work please see our Sprint announcement here. All of the outputs and educational materials will be available here.
As evidenced by the breadth of issue areas participants grappled with in their work, the responsible adoption of digital identification systems is not guaranteed. Careful consideration of risks, a commitment to inclusivity throughout the lifecycle of the technology, and robust safeguards against abuse are minimum requirements. Our discussions underscored the need to scrutinize urgent appeals to resolve longstanding social inequalities through technological bandaids.