Congressional Hearing 7/8/2020

Exposure Notification and Contact Tracing: How AI Helps Localities Reopen Safely and Researchers Find a Cure

On July 8th, 2020, a congressional hearing was held by the Task Force on Artificial Intelligence under the Financial Services Committee. The virtual meeting was entitled “Exposure Notification and Contact Tracing: How AI Helps Localities Reopen Safely and Researchers Find a Cure.” The purpose of the hearing was to explore the perceived trade-off between digital exposure notifications and contact tracing and consumer data protection and privacy. Opening remarks underscored the importance of maximizing life and economic liberty without sacrificing individual privacy in American society. Emphasis was brought onto the fact that for digital technologies to be truly effective, a strong participation from citizens is necessary and for them trust is imperative. To that end, four expert witnesses—Brian McClendon, Krutika Kuppalli, Andre M. Perry, and Ramesh Raskar—were called to testify.

Professor Ramesh Raskar, the founder of the PathCheck Foundation, condenses the Congressional issues into two fundamental questions: (1) how technology and technological innovation can help manage COVID-19 and (2) how the former can be accomplished while still preserving personal privacy, liberty, and freedom. Ultimately, bottom-up digital contact tracing can augment top-down manual contact tracing efforts significantly—this hybrid solution is a viable and scalable method to reduce the spread of the virus. Hear Ramesh’s testimony.

The bottom line is that even in the absence of social distancing and widespread testing, just 50% adoption of contact tracing can lower the spread factor RO to 0.5, eliminating the pandemic. While the witnesses acknowledged six main concerns with digital contact tracing (fraud, privacy, accuracy, adoption, equity, and engagement), they also provided robust recommendations for mitigating potential issues. It is critical that digital contact tracing be considered a serious measure for combating COVID-19.

Top Concerns from Congress with DCT

Recommendations

Fraud—People taking advantage of manual contact tracing, encouraging consumers to share personal or financial information without cause. Smartphone apps should have the proper privacy controls built-in to secure data, only to be shared with institutions authorized by the consumer. Similar to banking apps.

Privacy—Fear that consumer data collected without consent and/or sharing of data with third party data brokers. Need for transparency around data collection, storage, and use so that the individual is aware of every step. Technology must be developed by credible, non profit organizations who leverage open source solutions to provide transparency and are open for scrutiny by the public. Any data sharing is opt-in by the individual, anonymized, and only shared amongst those  in the public health workflow. (Specifically this is specific key codes shared by other phones to identify potential exposure to a positive COVID-19 case. No other information is collected or made available to the approved manual contact tracing partner.

Accuracy—Difficulties in pinpointing exact locations and/or who somebody has come in contact with. Also difficult to reach people without smartphones for high penetration. Use multiple methods that combine high-tech (GPS, WiFi, and Bluetooth) with low-tech (QR codes, card swiping, beacons and manual check-ins). Need for inclusive solutions.

Adoption—Concern was shared due to the lack of comprehensive national pandemic strategy leading to a lack of coordination between states to track the virus across borders. Many apps are not interoperable. This can lead to inaccurate real time data that is necessary to control the virus. Need for a National Pandemic Response Service via a national decentralized AI platform to coordinate data sharing and contact tracing alliances across states and counties, monitor current cases, provide insight into policy decisions, and predict future spread.

Equity—Need to ensure that DCT will not exacerbate existing structural health inequities that disproportionately affect vulnerable populations. Ensure that technologies (1) are built with input from, (2) receive feedback from, and (3) are tested by those vulnerable communities to ensure that AI biases do not amplify existing biases in the real world.

Engagement—Need to earn the trust and support of local communities. Also need to reach a high enough level of penetration for effective DCT. Target communities to (1) search for manual contact tracers who are trusted and knowledgeable and (2) create pockets of people with high levels of DCT penetration (i.e. universities, military bases, etc.)

Bottom-up digital contact tracing can augment top-down manual contact tracing efforts significantly—this hybrid solution is a viable and scalable method to reduce the spread of the virus. Pathcheck Foundation provides DCT and exposure notification solutions that are highly privacy-preserving and require very little personal information in order to increase adoption which in turn can save lives and help reopen economies without having to wait for cost-prohibitive testing infrastructure or scaling up manual contact tracing. Even in the absence of social distancing and widespread testing, just 50% adoption of contact tracing can lower the spread factor RO to 0.5, eliminating the pandemic. This adoption rate is possible given smartphone penetration in just the US is 90% among 18-64 year olds and it is this segment that are anxious for a solution to keep communities safe and avoid re-closing their communities due to a resurgence.