key: cord-0669195-d26ypb5a authors: Gupta, Abhishek title: The Gray Rhino of Pandemic Preparedness: Proactive digital, data, and organizational infrastructure to help humanity build resilience in the face of pandemics date: 2020-11-05 journal: nan DOI: nan sha: fb99caab40bc5979ff92097993eb511b566a3de9 doc_id: 669195 cord_uid: d26ypb5a COVID-19 has exposed glaring holes in our existing digital, data, and organizational practices. Researchers ensconced in epidemiological and human health work have repeatedly pointed out how urban encroachment, climate change, and other human-triggered activities and patterns are going to make zoonotic pandemics more frequent and commonplace. The Gray Rhino mindset provides a useful reframing (as opposed to viewing pandemics such as the current one as a Black Swan event) that can help us recover faster from these (increasingly) frequent occurrences and build resiliency in our digital, data, and organizational infrastructure. Mitigating the social and economic impacts of pandemics can be eased through building infrastructure that elucidate leading indicators via passive intelligence gathering so that responses to containing the spread of pandemics are not blanket measures; instead, they can be fine-grained allowing for more efficient utilization of scarce resources and minimizing disruption to our way of life. COVID-19 has exposed glaring holes in our existing digital , data , and organizational 1 2 3 practices. Researchers ensconced in epidemiological and human health work have repeatedly pointed out how urban encroachment, climate change, and other human-triggered activities and patterns are going to make zoonotic pandemics more frequent and commonplace . The Gray 4 Rhino mindset provides a useful reframing (as opposed to viewing pandemics such as the 5 current one as a Black Swan event) that can help us recover faster from these (increasingly) 6 frequent occurrences and build resiliency in our digital, data, and organizational infrastructure. Mitigating the social and economic impacts of pandemics can be eased through building infrastructure that elucidate leading indicators via passive intelligence gathering so that responses to containing the spread of pandemics are not blanket measures; instead, they can be fine-grained allowing for more efficient utilization of scarce resources and minimizing disruption to our way of life. Yet, pervasive monitoring poses significant privacy and trust challenges limiting the efficacy of 7 8 such measures. In addition, ad-hoc approaches further diminish trust in these measures. Participatory design and proactive development to generate preparedness will not only elicit 9 higher levels of trust, but also help to build infrastructures that are tailored better for the communities that they are meant to serve and help us better weather the harmful societal impacts of pandemics. From a digital infrastructure standpoint, continual investments in extending the reach of networks and internet access to the "last-mile" areas that are as of yet uncovered is important if we're to embark on helping all communities recover swiftly by sharing with them the necessary tools that allow them to function during periods of distancing and other protective measures mandates by health authorities. Sparsely connected regions have faced a disproportionate burden of this current pandemic due to their inability to access digital services, hindering their 10 ability to work remotely (to the extent that their workplaces allowed for that). Parallel investments in extending high-speed internet via existing wired and wireless networks can help reach sparsely populated regions . This will enable better data collection, leading 11 indicators generation, and tailoring policy responses to combat the pandemic in these regions. Reopening of the economy can also be done in an informed manner rather than the current trial-and-error approach that has led to recurring ebbs and flows of cases . 12 Here we approach the fundamental tension that has plagued the current technological deployment of solutions to combat COVID-19: privacy intrusions due to an expansion of surveillance infrastructure, whether benign or one with ulterior motives . There are three 13 14 measures that can be adopted to mitigate these concerns: transparency in the design, development, and deployment of contact-tracing and other technologies, use of privacy-preserving technologies like differential privacy (DP) as opposed to ad-hoc 15 anonymization and other unproven methods, and proactive and premeditated data trusts run by communities themselves. Adopting transparent design, development, and deployment of digital contact-tracing has been shown to be better received than taking a closed-door approach as seen in the face-off between PEPP-PT and DP-3T protocols in Europe. The Apple-Google toolkit when adopted by governments has also shown to be more effective in part (though potentially confounded by 18 higher interoperability and ease of use) because of the technical standards and considerations 19 being made public ahead of time. When the design choices are made public ahead of time, iterated open under public scrutiny, and have a tracked version history (via a service like Github) as was used by DP-3T , it engenders more trust from users and increases the adoption 20 of the technology bolstering its efficacy. Time and again , it has been demonstrated that ad-hoc anonymization doesn't offer the most 21 22 robust guarantees in terms of personal data protection. Yet, it is the dominant technique mentioned whenever discussions on privacy arise. Whether in popular media or in policy-makers' offices, a more thorough understanding of techniques like DP, which is able to provide mathematically verifiable bounds on the protection of personal data, will be essential. 23 More so, practical and highly public demonstrations using DP, for example the push from the US Census in 2020 , will showcase that it is not a theoretical construct and can be scaled to a 24 national level. Additionally, aligning language in privacy legislations and regulations that are closer to how such techniques might be realized in practice will help to strengthen the case for 25 adoption of these industry best practices. Ad-hoc creation of data trusts tends to create discord and distrust in the communities that these solutions are meant to serve . A priori creation of stewardship mechanisms for opportunity for engaged discussion arriving at agreements that are aligned with the values upheld by the communities . The creation of well-established playbooks and organizational measures such as dedicated bodies and committees that are composed of the necessary experts who have a long history of engagement on the technical, scientific, and social aspects of pandemic management helps to bake-in intelligence into the process through continual learning and iteration, something that is proving to be challenging as ad-hoc measures are springing up across the world . Unaware of 28 the work being done by others, and duplicating research and creating best practices from 29 scratch has led to uneven responses . In a pandemic, such responses are particularly harmful 30 since they limit the effectiveness of the measures undertaken by everyone else. Ultimately, embodying this Gray Rhino mindset has the benefit of ushering in an era of pandemic preparedness that taps into the best practices across digital, data, and organizational infrastructure that shifts the paradigm from being reactive to proactive. It is also centered on capacity building in a deliberate and continual manner to reduce uncertainty and the severity of negative consequences when a pandemic strikes. Respecting ethical considerations including privacy, data rights, context, and non-discrimination through the empowerment of people who are disproportionately affected by pandemics will help us utilize technology that truly does benefit everyone. DP-3T/documents Robust de-anonymization of large sparse datasets What the Surprising Failure of Data Anonymization Means for Law and Policy Formal verification of differential privacy for interactive systems Disclosure Avoidance and the 2020 Census Response to Office of the Privacy Commissioner of Canada Consultation Proposals pertaining to amendments to PIPEDA relative to Artificial Intelligence Report prepared by the Montreal AI Ethics Institute In Response to Mila's Proposal for a COVID-19 and the world of work Accelerating a safe and effective COVID-19 vaccine COVID-19 and the world of work