Open Questions on Good ID: What is Bad ID?
We aspire for all ID to be Good ID, but we also know that journey includes obstacles and requires more learning, given the nascent sector. We are committed to collaboration with other Good ID champions as well as constructive critics who want to help answer the difficult questions and find innovative solutions. In this new series, we will tackle open questions, one at a time, to share our perspective and invite others to offer theirs — all with the goal of adding more clarity on Good ID and finding the ideas that will move Good ID into practice.
In the second post in this series, we wanted to take the opportunity to start a conversation on bad ID.
Earlier this year, Omidyar Network published its point of view on Good ID. What defines a Good ID is its worthiness of the public’s trust. A Good ID unlocks opportunities fairly, enabling all of us to engage in society on our own terms and safeguard our identities against misuse. Good ID is distinguished by features like ardent transparency, accountability, privacy, inclusion, personal agency, and security. These features can manifest in different ways in various contexts and still achieve Good ID norms. However, when they are altogether missing, bad ID can develop.
Cases of exclusion, discrimination, surveillance, manipulation, and other forms of control are representative of bad ID and are unacceptable. A bad ID breeds distrust, facilitates control, and disempowers individuals and societies.
Whether by accident or design, identification issued by a government or business can result in bad ID scenarios. Bad ID can also result from the digital trails we leave behind when we interact with technology. Both formal and defacto ID create an indelible and rich image of who we are that is increasingly attractive to bad actors. Strong design and defenses are necessary to ensure all ID is, and remains, Good ID.
Omidyar Network and its partners are engaging both public- and private-sector decision makers to better understand and actively prevent these situations. Identity technology is relatively new, and the risks are unknown to many key decision makers. Identity systems require careful stewardship to ensure they remain a force for good and not a means of control.
Whether digital or analog, we believe all systems that interact with our identity data need to be regularly assessed for their potential to control and otherwise harm people. We are actively monitoring two risk areas that, if left unchecked, can nurture bad ID.
Business models and incentives
Certain business models can incentivize bad practices. For example, a company designing technology that gets smarter by collecting personally identifiable information is more likely to take advantage of people’s cognitive biases and less likely to secure consent, respect privacy, and return value to the user in the pursuit of growth. These conditions can lead to bad ID as well as to disinformation, algorithmic biases, addiction, surveillance, and inequality.
As an early-stage, impact investor, we take our role in advancing responsible technologies seriously. We are working with several other venture capitalists and startups to shift the current paradigm in the data economy. We support changing the incentives to spur more privacy-enhancing solutions, and leveraging regulation and competition as stimulus for innovation that benefits society. We want to encourage others to join this effort by investing in, designing, and testing responsible technologies and then sharing their success and impact.
Policy and design choices
Decisions that deprioritize transparency, accountability, privacy, inclusion, user value and control, and security, can result in bad ID. There are thousands of choices involved with an identity system, and we may not even realize these decisions have been embedded in policy, technology, and standard practice until we experience the negative impacts. To minimize bad ID scenarios, it is critical to evaluate upstream decisions for potential downstream harms. For example, could steep financial penalties for not registering for an ID coupled with tight enrollment deadlines disproportionately exclude and punish rural and low-income residents? And could loopholes in a system’s design lead to the leakage of someone’s health information, religion, ethnicity, or sexual orientation? Could these impacts further amplify social tensions, enable persecution and violence, assist censorship and control, motivate indiscriminate detainment and deportation, and facilitate transgenerational poverty?
We first wrote about the upsides and downsides of identification schemes in 2017. What we defined as bad identity then still applies—in both the digital and analog worlds—but we know it will need to continue to evolve as new harms are imagined and recognized in practice. How would you define bad ID today and in the future? What tools do decision makers need to spot bad ID choices, impacts, and outcomes, and recalibrate toward Good ID?
We would encourage Good ID champions, from all sectors of society, to have more conversation about the most consequential design choices and practices and to draft guidance that enables better immediate protections and long-term outcomes. Practical tools, like the Ethical OS Risk Mitigation Checklist, are in high demand by decision makers trying to navigate these choices well and deliver Good ID.
Research on the efficacy of specific incentives and choices is also needed. In the past three years, we have funded several studies of identity systems across different contexts, including the largest, objective annual household-level assessment of the Aadhaar system in India. Our civil society partners there have very effectively leveraged evidence of the system’s negative impacts to advocate for design limits and governance improvements.
To replicate this success around the world, we will continue to invest in responsible technologies as well as nonprofit organizations that are approaching these problems with urgency, remedies, and an eye on how Good ID can be part of the solution. Your support in generating more research, practical guidance, and other tools are important contributions in this shared mission.
Please help unlock the full potential of Good ID by sharing your learning, viewpoints, projects, events, and other resources on the Good ID online platform — www.good-id.org and @GoodID.
You can also read the first post in the “Open Questions on Good ID” series here.
In a forthcoming post, we will continue this conversation and explore the situations in which we would not recommend digital ID.