When COVID-19 emerged, many of us felt the instinct to use our technical skills to contribute something—and fast. But as researchers and technologists at Stanford, we also felt deep concern, having witnessed technologists’ blind spots and biases give birth to many dangerous technologies, including digital gaydar, deepfakes, discriminatory AI, AI surveillance and more. Even well-meant technologies can shift power away from those they purport to help. We have come to recognize that while the desire to help during COVID-19 is right, the rush to push just any COVID-19 technology is wrong and even has the potential to kill.
Trying to innovate their way back to normal, many technologists without previous medical or ethical expertise have proposed or deployed projects to calculate risk scores, trace contacts, model disease patterns and enforce quarantines. This inundation of ill-advised projects has led people to fall victim to misinformation and scams, eroded trust in science and provided cover for governments to expand their powers. Most new COVID-19 technologies risk adding to the chaos and eroding fundamental freedoms. With the stakes so high, here are four questions we call on ourselves and fellow technologists to answer before pushing technological responses to COVID-19:
1. Are you listening to experts and vulnerable communities?
Develop these relationships long before deployment to ensure you understand the social context, what will be helpful, and what will be harmful. If you are not already working with experts, start by finding credible sources of expert information. Meanwhile, while conversations and apps like Nextdoor and Twitter can surface the needs of some, these spaces often obscure the voices of the most vulnerable—including communities without access to technology; people who are unhoused, in nursing homes or in prisons; and those who cannot speak freely. Find people and organizations that center vulnerable communities. Listen carefully. What do they think is most pressing? Do they want you to build your technology for them, with them, or not at all?
2. Can you join existing efforts?
The process of open listening will lead to many calls to action. Investigate the solutions being called for. There are also groups such as U.S. Digital Response and Digital Aid, that match technologist volunteers with projects that would benefit from their skills. Search for a team that can speak to all four of these questions confidently and a project that has appropriate experts, community involvement, infrastructure and ethical frameworks in place.
3. Can your technology do what you say it’s going to do?
Will your solution improve outcomes in the real world, or might it only work in simulated environments? Do you claim your technology is scientific? If so, it must uphold scientific standards and avoid pseudoscientific approximations—including pseudoscientific machine-learned approximations. Can you complete and test your work at the desired scale in the desired timeframe? Down the line, will you have the resources to maintain your project, or might you abandon it? Document your sources of data, data preprocessing, modeling and analysis of results, making transparent the assumptions and limitations of your project. As an example, see this paper on the privacy implications of contact tracing and the authors’ explicit statement of how their ideas should and should not be used. In many cases, your technology’s limitations mean it should not influence policy decisions; state this up front and repeat it as necessary.
4. How does your technology shift power?
Finally, consider whom your project shifts power away from and whom it shifts power to. Ownership of data is a form of power: Do you provide meaningful opt-in to data collection? Whom are you giving access to this data? Do you inform users exactly what the service will and will not do and enforce those commitments with privacy-by-design principles and data governance? Many people are scared and willing to trust technology more than usual; we must hold ourselves accountable to this trust. Moreover, some solutions enable governments to expand mass surveillance or otherwise expand their power. This is especially insidious because, once created, government powers rarely go away. Reflect on who will have access to your technology and whether it will help vulnerable people or compound circumstances already stacked against them.
In our own work, we commit to reaching clarity on each of these four questions, and we encourage fellow technologists to do the same. If there is any question you cannot answer confidently, embrace your responsibility to bring your project to a close to avoid harming others. If we don’t contribute to the response to COVID-19 by coding, that’s okay. Here are ways we can help that don’t involve building new technologies:
This is an uncertain time, and many of us may feel drawn to the clarity of technological solutions. But as we strive to contribute to our communities, we must make sure we are practicing humility and thinking critically about our actions. With the stakes so high, this is no time to move fast and break things.
Read more about the coronavirus outbreak from Scientific American here, and read coverage from our international network of magazines here
eTurboNews: Seychelles has a population of slightly over 100,000 citizens. Nigeria has a population of…
Domestic and foreign tourism in Hungary is expected to increase by 10-15 percent compared to…
eTurboNews: Six people including Captain Chet Bahadur Gurung and five Mexican nationals were aboard the…
According to the Ministry of Trade, Industry, and Tourism's Directorate of Sector Analysis and Promotion,…
While the inflation rises and spreads to almost all consumer goods, 58% of French are…
Naturist travelers should be aware that many beaches in Portugal do not allow nudism. To…