Communications of the ACM
Closing the 'Trust Gap' in Contact Tracing/Exposure Notification Technology
Contact Tracing has been a valuable tool of public health for decades. Helping to eradicate smallpox and contain the spread of ebola, manual contact tracing has established itself as one of the tools for managing and controlling pandemics. Unsurprisingly, in our technology-centric world, as Covid-19 reduced the world to lockdown status, leading smartphone providers and app developers reached the conclusion that technology-enabled contact tracing would increase the efficiency and accuracy of contact tracing. However, the digitization of contact tracing also presents an array of challenges.
The most ubiquitous solution to date is from Google/Apple teaming to provide an exposure notification system API turning Bluetooth Low Energy (BLE) signals into a system to trace each of your contacts based on distance and time in contact. There are other examples around the world; some use BLE, others use GPS, and still others are examining solutions such as sonar. Each of these solutions is designed to replace and overcome one of the challenges of analog contact tracing, the fallibility of human memory, by rigorously tracing contacts prior to a positive test. Most of these solutions were produced very quickly—in some cases, source code appeared within weeks of the World Health Organization's call of Pandemic on March 11. That sort of speed has consequences for completeness, thoroughness, and diligence.
The notion of improving analog contact tracing is laudable, but the digital solution implementation has been far short of an obvious success. According to some experts, the adoption efficacy threshold for contact tracing to become effective must run above 60% of a given population. However, Oxford now claims that statistic was misunderstood. Highlighting the fact that this is all uncharted waters and the world is grasping for a map of some kind. Is there a threshold of effectiveness? How do we measure success?
In practice, Australian adoption has been decent, but now the app shows more 1-star ratings than any other, with claims of draining batteries, failed updates, interference with other Bluetooth devices, and just plain not working. In Europe, adoption rates remain low in Italy and France, while German's Corona-Warn app has been downloaded 16+ million times for about a 20% adoption rate. Is that good? Is that effective? In Ireland, by one measure there were a total of 58 users registered positive tests in the app's first three weeks of operation through to July 28, generating 137 close contact alerts. Of these, 129 opted to get a follow-up call from Ireland's contact tracing team. If one life is saved, is that effective? Maybe. Effectiveness measurement will require some more rigorous analysis. For now, the jury remains out.
In the U.S., where apps are proceeding more slowly, only 20% of total participants in a Brookings Institute survey said they would "strongly support" a contact tracing app, compared to 36% who "strongly opposed." Only 19% of respondents were "extremely likely" to download a hypothetical contact tracing app, while 42% said they were "extremely unlikely" to do so. It is clear there is a significant gap between realized adoption and comprehensive adoption. Trust remains a strong impediment in all jurisdictions.
While contact tracing is regarded as a valuable tool of public health, it can only achieve its function through the acceptance and efforts of the individual acting in a collective. For many, participating in contact tracing highlights a conflict between societal good and individual security, and thus operators of contact tracing must foster an overwhelmingly welcoming and trustworthy environment for the individual. Transparency honors the individual with knowledge, clarity, and assurance about the requirements and risks associated with their personal sacrifice, empowering them to willingly support the collective good. Transparency is foundational to building trust.
Contact tracing technology may distinguish itself as a valuable component in the ongoing global struggle with Covid-19, but the technology does create additional identifiable dangers and risks with regards to the collection, storage, and use of people's sensitive personal data, such as where they travel and with whom they come into contact. While some jurisdictions have passed admirable privacy-protecting legislation and others continue to push new legislation through, most have taken confusing stances, or none at all. "In short, right now, we cannot see the watchmen across the widening 'trust gap'."
Yet the pandemic rages onl and a different calculus has been added to the health risks. After months of lockdown and damage to the economy, leadership has begun to reassess all risks associated with Covid-19. Even the Centers for Diseash Control recently testified that kids needed to be back in school because of the health risks of NOT attending. Leaders around the world are keen to find ways to get people back into society, back to school, and back to work. Unemployed individuals and hand-to-mouth workers need to be on the job, and those who could work remotely still needed services like food and medical supplies. For employers, that means taking measures to make employees feel as safe as possible as they return to work. Along with ubiquitous screening and diagnostic testing, mask policies and social distancing, contact tracing is one of the measures that employers are considering to help manage these risks. This move by employers creates an additional challenge to the "trust gap." In a world of substantial unemployment, does "opt-in" have real meaning? Is there a genuine choice for an employee to use or not use digital contact tracing? This coercion makes the trust gap wider, and some companies do not have good track records with how they treat employees, let alone manage their sensitive personal data. Oversight and governance of employer-based contact tracing is vital for its success and to limit any downside risks.
Recognizing the need to bridge this "trust gap" between the traced and the contact tracing authorities, ForHumanity, a non-profit, endeavors to partner with contact tracing authorities in order to increase trustworthiness in digital contact tracing. We have proposed a series of solutions designed to increase the trustworthiness of contact tracing technologies and make them a paragon of public health's pandemic response. ForHumanity suggests three key areas where contact tracing authorities may be able to enhance their programs with additional best practices:
- Independent Governance
- Independent Audit (Best Practices in ethics, bias, privacy, trust and cybersecurity)
- Legislative and judicial remedies
Independent Governance
Independent means not influenced or controlled by others in matters of opinion, conduct, etc.; thinking or acting for oneself.
Governance means supervision; watchful care, overseeing.
Said plainly, this would be a panel of experts who are unaffiliated with both the authority permitting the contact tracing as well as the firm or organization implementing the contact tracing system. They should not be selected by the technology firm, nor should specific members be selected by the funding entity. This governance entity exists to provide oversight for the people being traced. Examples of best practice governance include:
- Transparency
- Best practice cybersecurity
- Layers of Opt-In/Voluntary participation
- Minimizing data collection, maximizing privacy
- Unbiased
- Ethical operations
- Trustworthy disclosure, accessible technology
However, ForHumanity's assistance is not a watchdog effort ready to pounce on the slightest misstep. Rather, Independent Governance is designed to build trust. The Board works with the contact tracing authorities to "get it right." Bringing thought leaders in the areas of ethics, bias, privacy, trust, cybersecurity, and the law, a ForHumanity Board of Governance has the global perspective and expertise to lend valuable insight in the design and maintenance of a robust contact tracing technology.
Independent Audit
This comprehensive approach to contact tracing will provide an Independent Board of Governance the tools and information it requires to fairly and responsibly oversee and advise the contact tracing program. We anticipate this comprehensive oversight will increase trust in contact tracing technologies and provide assurance to the public broadly. The audit promotes transparency, compliance, disclosure, and accountability.
With constant review of best practices, the Independent Audit of Contact Tracing is maintained by the ForHumanity Fellows who are scouring best-practices in contact tracing technologies around the world.
This team of experts have already made revisions to the Audit to accommodate recent court rulings on data privacy, like the EU High Court's ruling now commonly known as "Schrems 2." Best practices and new technologies, like sonar from SonarX and others, are being reviewed constantly. Likewise, since all digital contact tracing is new, the Fellows continue to monitor implemented digital contact tracing for effectiveness or ineffectiveness. A current example of potential ineffectiveness is BLE. BLE appears to be too unreliable (imprecise proximity, an ability to register contact through walls, and cybersecurity concerns) which may lead to too many false positives. Unfortunately, too many false positives and false negatives increases the "trust gap." The Fellows are constantly monitoring the world for best practices in digital contact tracing as a further means to bridging the "trust gap."
The Audit is a combination of information gathering and best practices. Many will be unable to meet the highest standard of compliance, while others will find their systems are mostly compliant but need some tweaking to maximize the benefit from the audit and achieve the highest level of trustworthiness.
Audit rules and standards are open and transparent to those who want to examine, review, critique, and enhance the body of work. This iterative process, combined with new Fellows joining ForHumanity regularly, ensure that the rules and standards are of the highest quality and always comprehensive and current.
Legislative and judicial remedies
Rather than go into these at length, it is easier and more effective to point to the piece that Mark Potkewitz, ForHumanity Fellow and Director of the Legal Innovation Centre at Ulster University, and I wrote in Legal Business World "Suggested legislative measures to overcome the contact tracing 'trust gap'." However, We can provide a bullet point summary of the suggestions here:
- Explicit restrictions on how government may use contact tracing data.
- Suppression remedy for data improperly shared with law enforcement.
- A Private cause of action against contact tracing entities and service providers with a negligence standard, accompanied by statutory damages.
- Statutory right to a data subject access request.
- Clear and controlling legislation or regulation which states how contact tracing data will be used in any aggregate or anonymized form with associated consent.
- Criminal or civil punishments or sanctions for misuse or improper sharing of contact tracing data with an appropriate mens rea standard.
- Criminal or civil punishments or sanctions for knowing misuse of a contact tracing solution or knowingly reporting false information.
- Legislation which directly suspends any data retention rules and guarantees deletion of data to the highest extent.
- Non-discrimination and interoperability across jurisdictions.
The point of these legislative and judicial remedies is to ensure that contact tracing/exposure notification technology is safeguarded from being maligned, mistrusted, or abused by authorities through mission creep, misappropriation of data, or a failure to keep promises. These legislative best practices would make misuse extremely difficult. If comprehensively implemented as above, then contact tracing technology would be "wrapped in bubble wrap with belts and suspenders," designed to protect it exclusively for maintaining public health. The arguments are simple, authorities have many ways to gather data, many ways to surveil its citizenry, using contact tracing simply renders this public health tool useless.
In conclusion, contact tracing has gone digital. Implementations are occurring throughout the world from jurisdictions as large as nation-states down to localized solutions as small as for restaurants. These solutions are one part of a comprehensive risk management solution designed to get people back to normalcy. They can be very effective tools for slowing the spread of a highly infectious disease, but we must remember that they are new, hastily rolled out, and predicated on the public’s trust. We must endeavor to do all that we can to grow the trust in these technologies, and that requires the operators of these technologies to humbly submit to best practice standards of transparency and accountability. ForHumanity stands ready to help.
Guest blogger Ryan Carrier is executive director of ForHumanity, a non-profit organization created to examine and mitigate the downside risks associated with Artificial Intelligence and Automation. Independent Audit of AI Systems is one such risk mitigation tool.
No entries found