ImPACT 2020: PACT Project Holds First Collaborative, Virtual Workshop to Tackle Key Technical Issues in Private Automated Contact Tracing

2020-04-16 - 11 minutes read

Today, during the first ImPACT 2020 event, world-leading computer scientists in cryptography, cybersecurity, and wireless networking got together with top public health, privacy, and legal experts to discuss contact tracing. Contact tracing, the process of finding out who might have been exposed to a virus that is spread through communities, is crucial to containing COVID-19. Both manual contact tracing and private automated contact tracing will be critical in the days ahead for not only individual states or nations, but also for countries to exit existing self-imposed restrictions. As Massachusetts Governor Charlie Baker stated during his remarks, “[t]his work you’re doing today is incredibly important. We’re glad to have people there who can benefit from the shared knowledge that will be brought forth during these conversations.”

The conference was sponsored by the PACT project, which came together just three weeks ago. This project involves shaping “the extraordinarily powerful technical tools at our disposal, in service of public health goals, with strong privacy protections” noted Daniel Weitzner, the founding director of MIT IPRI and Co-PI of the PACT project. MIT Professor and other Co-PI of the project, Ronald Rivest, later emphasized this final point. Noting that  “privacy is a key goal here… We want to figure out what’s relevant for medical purposes and that only, no more.”

A full recording of this event will be available soon and linked to on this page, but in the meantime, here’s a quick recap of some of the discussions during the meeting. 

If you are interested in learning more about ImPACT 2020, please reach out to Impact2020@mit.edu.

The Importance of Contact Tracing and Preserving Privacy

Daniel Weitzner kicked off ImPACT 2020 by discussing the importance of this project and the need for it to be integrated with existing methods and processes. “From the perspective of the PACT project, it’s clear that whatever is done with smartphone technology should be a complement to the core public health and manual contact tracing strategies now being implemented” Weitzner noted.

Weitzner was followed by a video message from Commonwealth of Massachusetts Governor, Charlie Baker.

Governor Baker spoke openly about the importance of contact tracing. As part of these remarks, he noted that “[t]his issue about contact tracing, particularly on times like this with a pandemic like the one we’re all dealing with, is not just an issue for Massachusetts, it’s not just an issue for the country, it’s, frankly, an issue for the world.” He also noted how important contact tracing was to both trace the virus as well as give people confidence over time that they can safely return to work. He also emphasized the importance of privacy for projects like this.

Governor Baker ended his remarks by stating, “[a]nd I just want you to know, as the Governor of the Commonwealth of Massachusetts, I happen to think this is one of the most important initiatives we’ll be pursuing in our efforts to keep people safe and provide them with confidence that we can get past this and on to whatever the next new normal looks like. Good luck and God bless.”

Ronald Rivest spoke next and shared an overview of the meeting. During his introduction to ImPACT 2020, Rivest touched on three main points the event would cover. These include:

  • Can automatic contact tracing be a useful supplement to manual contact tracing?
  • How can smartphones use Bluetooth to see if people are close enough for infectious coronavirus spread?
  • How can automated contact tracing be done in a privacy-preserving manner?

Public Health Requirements for Contact Tracing

For the first discussion of the event, Louise Ivers (MD, MPH, DTM&H, Executive Director, MGH Center for Global Health, and Associate Professor at Harvard Medical School) led a talk that included Louis Gutierrez (Executive Director of the Massachusetts Health Connector) and KJ Seung (Chief of Policy and Strategy for Massachusetts COVID19 Response, Partners In Health).

A few of the key points in this discussion were:

  • Contact tracing is nothing new, and has been used for a while to stop the spread of infectious diseases. 
  • In Massachusetts, there are around 1,000 people working together to perform manual contact tracing via a Salesforce platform. This already is a technologically complex situation to handle.
  • There are a few problems with adding technology into this scenario. One of which is that not everyone has a phone and not every person with a phone is reachable. As such, people will need to work together and use different techniques to reach the greatest number of people possible.
  • We need an approach that can be uniformly rolled out, and if participation is voluntary, it needs to be easy.
  • Participants noted that they would welcome new technologies that could handle the “easy cases” and remove the load from their small team so they could focus on the most-complicated cases. 

Investigating Bluetooth Sensor Engineering

In the next session, ImPACT speakers covered the topic of Bluetooth sensor engineering. This talk was moderated by John Wilkinson (leader of the Cyber System Assessments Group at MIT Lincoln Laboratories) and included remarks from Swarun Kumar (Assistant professor at CMU), Jennifer Watson (Assistant group leader, MIT Lincoln Laboratories), Han Yang Tan (Singapore Government Technology Agency), Joel Linsky (VP of Technology, Qualcomm), Douglas Reynolds (Senior Member MIT Lincoln Laboratories), and Thomas Wiegand (Executive Director, Fraunhofer HHI). 

A few of the topics of discussion during this panel were:

  • Plotting a contact tracing ROC curve and determining what combination of range and exposure time would lead to a user receiving an alert, as seen in the following screenshots from Wilkinson’s talk:

  • Why proximity sensing is hard, an example of which is seen in the following screenshot from Kumar’s talk, and the reasons we are using Bluetooth for automatic contact tracing.

  • Potential detection strategies and what successful detection relies on.
  • Discussing an app currently in use with 1 million users.
  • Bluetooth low-energy advertising and pathloss estimations.
  • The need for task definition, metrics to know if it apps like these are working, and data.
  • An existing proximity tracing app that considers public health policy and aims to apply an ethical approach.

Discussing Privacy and Cryptographic Protocols

In the last panel of the day, we delved more into PACT, as well as privacy and cryptographic protocols. This panel was moderated by Emily Shen, the technical lead of PACT from MIT Lincoln Laboratories. Also speaking in this panel were: Stefano Tessaro (Associate Professor at UW), Carmela Troncoso (Assistant Professor EPFL), Henry de Valence (Zcash Foundation / TCN Coalition), and Ran Canetti (professor of Computer Science at BU).

A few of the topics discussed in this panel include:

  • How the community can move forward quickly and productively and identify paths toward convergence.
  • Focusing on preserving privacy in decentralized architecture. 
  • A summary of PACT, as seen in the following screenshot from Tessaro’s presentation:

  • Keeping the main philosophy of do no harm in mind via purpose limitation and data minimization.
  • A look into the TCN coalition and its efforts to collaborate between different applications.
  • Mitigating COVID-19 via contact tracing without long-term harm to civil liberty, as seen in the following screenshot from Canetti’s presentation.

What’s Next?

Weitzner and Rivest ended the meeting by looking ahead, where there are certainly challenges, but also opportunities for collaboration and progress. 

For instance, to move forward, we need to look into best practices and understand privacy on a broad social systems level. Further, there is a need for automatic contact tracing to integrate seamlessly with manual contact tracing, as well as cross-platform convergence and interoperability, at least regionally. As this is a new design pattern, there are bound to be vulnerabilities, so open-source source code and operational transparency are important as this project progresses. We will discuss more about PACT and the challenges and opportunities this project presents in future meetings.