‘Colorblind’ Tech is Killing Us: Why COVID-19 Tech Must Focus on Equity
By Mary Gray, Senior Principal Researcher at Microsoft Research, an E.J. Safra Center for Ethics Fellow and Faculty Associate at the Berkman Klein Center
This essay is part of a collection written by members of the Berkman Klein Center’s Working Group on Digital Pandemic Response. The group, made up of experts from academia, civil society, the public sector, and industry, takes on difficult questions around the use of digital tools and data to help attenuate the COVID-19 pandemic. Each essay is the perspective of the author, not of the Berkman Klein Center.
COVID-19 has taken the lives of more than 120,000 people in the United States and debilitated millions more since arriving on our shores. While anyone can contract the virus, there is a pattern to who gets sick and dies from it here. Yet very few of the Tech Sector’s responses to the disease seem to acknowledge — let alone respond to — this fact.
In some ways, this is not surprising. Tech prides itself on achieving scale by being agnostic to context, instead tapping into some universal human need. And, for the first few weeks of national news coverage, COVID-19 seemed to have no bounds. Stories focused on cruise ship passengers, like those on the Grand Princess anchored off California’s coast, and nursing home residents, like those devastated by COVID-19 in Washington State in late February. Of course, COVID-19 can silently strike anyone, anywhere. But, as early as April, news outlets, like the Washington Post, confirmed what many working long hours in our country’s hospital emergency wards already suspected: COVID-19 was infecting and killing specific groups of Americans at alarming rates.
As states began to more systematically document the demographics of those falling ill, it quickly became clear that Black and Latinx communities were far more likely to suffer the lethal impact of the virus. This June, the Boston Globe reviewed new data released by Massachusetts detailing the sharp racial divide in that state:
Hispanic residents make up 12 percent of the population, but their rate of positive cases is nearly 30 percent. Similarly, Black residents represent about 7 percent of the population, but their rate of positive COVID-19 cases is double that, at 14.4 percent. Overall, the state data show, the rate of positive cases among Black and brown residents is more than three times that of white residents.
But, today, most technological innovations remain strangely ‘colorblind’ to the reality that racial inequalities play a significant role in where COVID-19 makes the most significant impact. Despite overwhelming data telling us that the most vulnerable, historically marginalized communities disproportionately bear the heavy costs of COVID-19, tech efforts still focus on an efficient moment of data collection and information-sharing — a ping or a text notice — designed to quietly let Person X (let’s call her Alice) know that they may have crossed paths with Person Y (let’s call him Bob) with COVID-19.
At a recent showcase of COVID-19 ‘Return to Work’ technologies, only two of more than a dozen demos include any health guidance with the alert that they’ve been exposed to COVID-19. The underlying assumption seemed to be that the most valuable role for tech was relaying individual test results or geolocation data. But to make a dent in COVID-19, tech will need to prioritize more than efficiency. It will need to think about and build for racial equity, too.
To be sure, there are good arguments for digital tracing and proximity notification applications deployed via smart devices to convey information. If just a portion of a population — say at a worksite or college campus — adopt these digital apps, we could reduce some of the work currently spent figuring if, where, or when someone crosses paths with the virus. But what is wrong-headed about this singular approach is not just the most recent string of studies suggesting that many in the United States are unwilling to download apps and share their health data. The bigger issue is that identifying who might be sick is meaningless if this key pillar of the epidemiological practice referred to as ‘contact tracing’ is not linked to tech-savvy systems for patient monitoring and care for those infected with COVID-19. It is like building a 911-emergency system to log calls for help disconnected from systems that ensure a first responder shows up to assist.
Put simply, today’s tech approaches to COVID-19 exacerbate the systemic racism and health disparities that have given the pandemic its grotesque shape in our country — because they ignore them. Tech built to quickly sort COVID-19 exposure assumes that the most efficient way to ferret out the virus is to start with specific, yet abstract, technical questions: How do we maintain both privacy and accuracy in data collection? But in the rush to solve this wicked hard tech problem of identifying who is sick in a privacy-preserving way, the Tech Sector has all but ignored building the technologies needed most: Systems to support the healthcare work of routing and managing care to those struck down by the virus.
For example, in preliminary work with colleagues at IDEO.org and Duke Health listening to and learning from Duke Health’s Pandemic Response Network, we see the need for tech that seamlessly routes patients to a team of other care workers who have the language skills and community ties to talk someone through self-quarantining and accessing local help. This is not a technically hard problem, but it will take a sophisticated human-centered technical solution to get the collaboration tools and information-sharing right. Ultimately, the most efficient tech efforts to combat the pandemic will be the ones that prioritize equity. Failing to focus on caring for communities of color and essential workers exposed to COVID-19 because of their occupations misses the most effective way to truly corner COVID-19.