Stefaan Verhulst

Letter to the Editor of Nature by Vittoria Colizza et al: “Digital contact tracing is a public-health intervention. Real-time monitoring and evaluation of the effectiveness of app-based contact tracing is key for improvement and public trust.

SARS-CoV-2 is likely to become endemic in many parts of the world, and there is still no certainty about how quickly vaccination will become available or how long its protection will last. For the foreseeable future, most countries will rely on a combination of various measures, including vaccination, social distancing, mask wearing and contact tracing.

Digital contact tracing via smartphone apps was established as a new public-health intervention in many countries in 2020. Most of these apps are now at a stage at which they need to be evaluated as public-health tools. We present here five key epidemiological and public-health requirements for COVID-19 contact-tracing apps and their evaluation.

1. Integration with local health policy. App notifications should be consistent with local health policies. The app should be integrated into access to testing, medical care and advice on isolation, and should work in conjunction with conventional contact tracing where available¹. Apps should be interoperable across countries, as envisaged by the European Commission’s eHealth Network.

2. High user uptake and adherence. Contact-tracing apps can reduce transmission at low levels of uptake, including for those without smartphones². However, large numbers of users increase effectiveness^3,4. An effective communication strategy that explains the apps’ role and addresses privacy concerns is essential for increasing adoption⁵. Design, implementation and deployment should make the apps accessible to harder-to-reach communities. Adherence to quarantine should be encouraged and supported.

3. Quarantine infectious people as accurately as possible. The purpose of contact tracing is to quarantine as many potentially infectious people as possible, but to minimize the time spent in quarantine by uninfected people. To achieve optimal performance, apps’ algorithms must be ‘tunable’, to adjust to the epidemic as it evolves⁶.

4. Rapid notification. The time between the onset of symptoms in an index case and the quarantine of their contacts is of key importance in COVID-19 contact tracing^7,8. Where a design feature introduces a delay, it needs to be outweighed by gains in, for example, specificity, uptake or adherence. If the delays exceed the period during which most contacts transmit the disease, the app will fail to reduce transmission.

5. Ability to evaluate effectiveness transparently. The public must be provided with evidence that notifications are based on the best available data. The tracing algorithm should therefore be transparent, auditable, under oversight and subject to review. Aggregated data (not linked to individual people) are essential for evaluation of and improvement in the performance of the app. Data on local uptake at a sufficiently coarse-grained spatial resolution are equally key. As apps in Europe do not ‘geolocate’ people, this additional information can be provided by the user or through surveys. Real-time monitoring should be performed whenever possible….(More)”.

Thea Snow at LSE Blog: “Sometimes, you learn about an idea that really sticks with you. This happened to me recently when I learnt about “legibility” — a concept which James C Scott introduces in his book Seeing like a State.

Just last week, I was involved in two conversations which highlighted how pervasive the logic of legibility continues to be in influencing how governments think and act. But first, what is legibility?

Defining Legibility

Legibility describes the very human tendency to simplify complex systems in order to exert control over them.

In this blog, Venkatesh Rao offers a recipe for legibility:

• Look at a complex and confusing reality…
• Fail to understand all the subtleties of how the complex reality works
• Attribute that failure to the irrationality of what you are looking at, rather than your own limitations
• Come up with an idealized blank-slate vision of what that reality ought to look like
• Argue that the relative simplicity and platonic orderliness of the vision represents rationality
• Use power to impose that vision, by demolishing the old reality if necessary.

Rao explains: “The big mistake in this pattern of failure is projecting your subjective lack of comprehension onto the object you are looking at, as “irrationality.” We make this mistake because we are tempted by a desire for legibility.”

Scott uses modern forestry practices as an example of the practice of legibility. Hundreds of years ago, forests acted as many things — they were places people harvested wood, but also places where locals went foraging and hunting, as well as an ecosystem for animals and plants. According to the logic of scientific forestry practices, forests would be much more valuable if they just produced timber. To achieve this, they had to be made legible.

So, modern agriculturalists decided to clear cut forest, and plant perfectly straight rows of a particular species of fast-growing trees. It was assumed this would be more efficient. Planting just one species meant the quality of timber would be predictable. In addition, the straight rows would make it easy to know exactly how much timber was there, and would mean timber production could be easily monitored and controlled.

 Reproduced from https://www.ribbonfarm.com/2010/07/26/a-big-little-idea-called-legibility/

For the first generation of trees, the agriculturalists achieved higher yields, and there was much celebration and self-congratulation. But, after about a century, the problems of the ecosystem collapse started to reveal themselves. In imposing a logic of order and control, scientific forestry destroyed the complex, invisible, and unknowable network of relationships between plants, animals and people, which are necessary for a forest to thrive.

After a century it became apparent that relationships between plants and animals were so distorted that pests were destroying crops. The nutrient balance of the soil was disrupted. And after the first generation of trees, the forest was not thriving at all….(More)”.

Paper by Angela Daly, S Kate Devitt, and Monique Mann: “In this chapter we argue that discourses on AI must transcend the language of ‘ethics’ and engage with power and political economy in order to constitute ‘Good Data’. In particular, we must move beyond the depoliticised language of ‘ethics’ currently deployed (Wagner 2018) in determining whether AI is ‘good’ given the limitations of ethics as a frame through which AI issues can be viewed. In order to circumvent these limits, we use instead the language and conceptualisation of ‘Good Data’, as a more expansive term to elucidate the values, rights and interests at stake when it comes to AI’s development and deployment, as well as that of other digital technologies.

Good Data considerations move beyond recurring themes of data protection/privacy and the FAT (fairness, transparency and accountability) movement to include explicit political economy critiques of power. Instead of yet more ethics principles (that tend to say the same or similar things anyway), we offer four ‘pillars’ on which Good Data AI can be built: community, rights, usability and politics. Overall we view AI’s ‘goodness’ as an explicly political (economy) question of power and one which is always related to the degree which AI is created and used to increase the wellbeing of society and especially to increase the power of the most marginalized and disenfranchised. We offer recommendations and remedies towards implementing ‘better’ approaches towards AI. Our strategies enable a different (but complementary) kind of evaluation of AI as part of the broader socio-technical systems in which AI is built and deployed….(More)”.

Paper by Adriano Koshiyama: “Business reliance on algorithms are becoming ubiquitous, and companies are increasingly concerned about their algorithms causing major financial or reputational damage. High-profile cases include VW’s Dieselgate scandal with fines worth of $34.69B, Knight Capital’s bankruptcy (~$450M) by a glitch in its algorithmic trading system, and Amazon’s AI recruiting tool being scrapped after showing bias against women. In response, governments are legislating and imposing bans, regulators fining companies, and the Judiciary discussing potentially making algorithms artificial “persons” in Law.

Soon there will be ‘billions’ of algorithms making decisions with minimal human intervention; from autonomous vehicles and finance, to medical treatment, employment, and legal decisions. Indeed, scaling to problems beyond the human is a major point of using such algorithms in the first place. As with Financial Audit, governments, business and society will require Algorithm Audit; formal assurance that algorithms are legal, ethical and safe. A new industry is envisaged: Auditing and Assurance of Algorithms (cf. Data privacy), with the remit to professionalize and industrialize AI, ML and associated algorithms.

The stakeholders range from those working on policy and regulation, to industry practitioners and developers. We also anticipate the nature and scope of the auditing levels and framework presented will inform those interested in systems of governance and compliance to regulation/standards. Our goal in this paper is to survey the key areas necessary to perform auditing and assurance, and instigate the debate in this novel area of research and practice….(More)”.

Paper by Paul Schmitt and Barath Raghavan: “To receive service in today’s cellular architecture, phones uniquely identify themselves to towers and thus to operators. This is now a cause of major privacy violations, as operators sell and leak identity and location data of hundreds of millions of mobile users. In this paper, we take an end-to-end perspective on the cellular architecture and find key points of decoupling that enable us to protect user identity and location privacy with no changes to physical infrastructure, no added latency, and no requirement of direct cooperation from existing operators. We describe Pretty Good Phone Privacy (PGPP) and demonstrate how our modified backend stack (NGC) works with real phones to provide ordinary yet privacy-preserving connectivity. We explore inherent privacy and efficiency tradeoffs in a simulation of a large metropolitan region. We show how PGPP maintains today’s control overheads while significantly improving user identity and location privacy…(More)”.

Fast Company: “Before he cofounded ride-sharing company Lyft, CEO Logan Green learned the intricacies of public transportation as a director on the Santa Barbara Metropolitan Transit District board. Venture capitalist Bradley Tusk worked as a communications director for Sen. Chuck Schumer and served as deputy governor of Illinois. Aerospace engineer Aisha Bowe says her six years working at NASA were “instrumental” to founding STEMBoard, a tech company that serves government and private-sector clients.

For these business leaders, “government service” isn’t a pejorative. Their work in the public sector helped shape their entrepreneurial journeys. And many executives from Silicon Valley to Wall Street have served at the highest levels in government, including Mike Bloomberg (a three-term mayor of New York), Megan Smith (former Google executive who served as Chief Technology Officer of the United States), and Rhode Island Gov. Gina Raimondo, a former venture investor  nominated to be U.S. Secretary of Commerce.

Today Fast Company is launching an initiative called Fast Government that aims to examine the cross-pollination of talent and innovative ideas between the public and private sectors. It is a home for stories about leaders who are bringing entrepreneurial zeal to state, federal, and local agencies and offices. This section will also explore the ways government service helped shape the careers of business leaders at some of the world’s most innovative companies.

As Sean McManus and Brett Dobbs explain in this accompanying piece, the talent pipeline in government needs refreshing. A third of federal civilian employees are slated to retire in the next five years, and fewer than 6% are under the age of 30. Young leaders, especially purpose-driven individuals looking to make a difference, might perhaps want to consider a stint in fast government….(More)”.

About: “Connected papers is a unique, visual tool to help researchers and applied scientists find and explore papers relevant to their field of work.

How does it work?

• To create each graph, we analyze an order of ~50,000 papers and select the few dozen with the strongest connections to the origin paper.
• In the graph, papers are arranged according to their similarity. That means that even papers that do not directly cite each other can be strongly connected and very closely positioned. Connected Papers is not a citation tree.
• Our similarity metric is based on the concepts of Co-citation and Bibliographic Coupling. According to this measure, two papers that have highly overlapping citations and references are presumed to have a higher chance of treating a related subject matter.
• Our algorithm then builds a Force Directed Graph to distribute the papers in a way that visually clusters similar papers together and pushes less similar papers away from each other. Upon node selection we highlight the shortest path from each node to the origin paper in similarity space.
• Our database is connected to the Semantic Scholar Paper Corpus (licensed under ODC-BY). Their team has done an amazing job of compiling hundreds of millions of published papers across many scientific fields.…(More)”.

Book by Aileen Nielsen: “Fairness is becoming a paramount consideration for data scientists. Mounting evidence indicates that the widespread deployment of machine learning and AI in business and government is reproducing the same biases we’re trying to fight in the real world. But what does fairness mean when it comes to code? This practical book covers basic concerns related to data security and privacy to help data and AI professionals use code that’s fair and free of bias.

Many realistic best practices are emerging at all steps along the data pipeline today, from data selection and preprocessing to closed model audits. Author Aileen Nielsen guides you through technical, legal, and ethical aspects of making code fair and secure, while highlighting up-to-date academic research and ongoing legal developments related to fairness and algorithms.

• Identify potential bias and discrimination in data science models
• Use preventive measures to minimize bias when developing data modeling pipelines
• Understand what data pipeline components implicate security and privacy concerns
• Write data processing and modeling code that implements best practices for fairness
• Recognize the complex interrelationships between fairness, privacy, and data security created by the use of machine learning models
• Apply normative and legal concepts relevant to evaluating the fairness of machine learning models…(More)”.

Paper by James Pow: “There are two important dimensions to the membership of mini-publics that are distinct from the membership of conventional representative institutions: the selection mechanism (sortition) and the profile of the body’s eligible membership (‘ordinary’ citizens). This article examines the effects of these design features on perceived legitimacy. A survey experiment in the deeply divided context of Northern Ireland finds no evidence that variation in mini-public selection features has an overall effect on perceived legitimacy, but there are important individual-level differences….(More)”.

Peter Addo and Stefaan G. Verhulst in The Conversation: “….This month, the GovLab and the French Development Agency (AFD) released a report looking at precisely these possibilities. “Emerging Uses of Technology for Development: A New Intelligence Paradigm” examines how development practitioners are experimenting with emerging forms of technology to advance development goals. It considers when practitioners might turn to these tools and provides some recommendations to guide their application.

Broadly, the report concludes that experiments with new technologies in development have produced value and offer opportunities for progress. These technologies – which include data intelligence, artificial intelligence, collective intelligence, and embodied intelligence tools – are associated with different prospective benefits and risks. It is essential they be informed by design principles and practical considerations.

Four intelligences

The report derives its conclusions from an analysis of dozens of projects around Africa, including Senegal, Tanzania, Uganda. Linking practice and theory, this approach allows us to construct a conceptual framework that helps development practitioners allocate resources and make key decisions based on their specific circumstances. We call this framework the “four intelligences” paradigm; it offers a way to make sense of how new and emerging technologies intersect with the development field….(More)” (Full Report).