By Alex Gekker & Anat Ben-David
With the rapid unfolding of the COVID-19 global pandemic, Israel was one of the first states outside East-Asia to impose involuntary surveillance measures as a means to combat the virus. Crucially, the government utilized the country’s permanent state of exception to bypass the parliament and deploy a hitherto classified anti-terrorism tool developed by its internal security service (Shin Bet) to track the location of coronavirus patients, identify infection-chains and notify citizens who have been in close proximity to an identified patient to self-quarantine. Despite the marked similarity to the Snowden revelations in terms of scope and granularity of data available to secret services on individuals, the extreme measures undertaken by the Israeli government were met by a legal battle ensued by a small group of activists and civil society organizations, but not by a public outcry. Rather, the majority of Israelis were willing to compromise their right to privacy for the technological protection offered against the virus, and expressed high levels of trust in the Shin Bet, even as the latter was often reluctant to take up the mantle. In this essay we draw on historian Daniel Rosenberg’s notion of “data before the fact” to reflect on how various uses of (big) data in Israel have led to compliance and confusion for the people involved.
Rosenberg suggests that data came to be a historically recognised category that is “given” (“data” means “that which is given” in Latin and thus not questioned or interrogated. Only the results coming from the data are. Recently, with the rise of massive data collection and machine learning techniques, data has further changed meaning, but retained that historically grounded sense of objectivity. This joins the tendency of computers being seen as “accurate” and “unbiased”— consider how Facebook claims that no user privacy is breeched because no humans are involved in seeing one’s personal details. Wendy Hui Kyong Chun and David Berry suggest that computer software becomes ingrained in the very language and metaphors of how we think about “thinking” today. Various data dashboards, including in combating Corona, cement the data-reflected reality rather than being tools for discussing alternatives. In the Israeli case, the discussion of possible responses to COVID-19 and their implications became entangled with the computerised data gathered on the disease’s spread, in a way that limited potential objections to the measures imposed. We show this across three distinct episodes.
Shin Bet Surveillance
After a publication by investigative journalists, the Shin Bet’s surveillance system was exposed. Called simply “the Tool” it has been in operation since 2002 and used for continuous trawling collection of all available cell-phone data from every mobile device in Israel and the Palestinian Territories. Officially used for counter-terrorism and previously (officially) used only targeted surveillance on specific individuals and in relation to a case, the Tool has nonetheless been employed for digital contact tracing across the entire land and marketed as an emergency extreme measure that is a “magic bullet” solution. Specifically the “Tool” allowed avoiding the need to deploy a voluntary digital contact tracing apps as done by other democratic states. Nonetheless such a civilian, Bluetooth-based phone app was developed, failed and re-developed again to a minimal reception by the Ministry of Health. One of the central arguments against adopting this latter, privacy preserving option, was due to the fact that the Shin Bet is collecting all data anyway. Yet, being a secret government agency, the Shin Bet is reluctant to share the specifics of the data collected. Early reports indicate a 5% wide margin of error. Out of 71 thousand people required to self-quarantine based on the tool’s data in the first week of July alone, about 22 thousand appealed, claiming to be false positives, and 60% of those appeals were accepted. Overall, the data presented of the effectives of the tool was lacking, included repeated numbers in different categories, and was unreliable. Nonetheless, the appearance of efficiency supported by seemingly impressive numerical data has led to a continuing adaption of the tool instead of other alternatives.
National Compliance Index
Another use of numerical data to create compliance rather than support policy was through the deployment of a “national index”, in cooperation with renowned behavioral economist Dan Arieli’s Kayma company. The index was developed by Kayma as a single entrant to an urgent tandem presented by Israel’s National Corona Response Centre. Despite potential financial and practical concerns, the company was selected to monitor various “commercial and civilian data sources” in order to track how compliant the population is with Covid regulations. Prominent on various platforms—including on the main page of the country’s most-read news website—and asking citizens to self-report on “compliance” such as hand-washing or mask-wearing, the index generates a variety of dashboard statistics, while being extremely opaque in its data sources. As in the previous case, the numerical data, information visualizations and dashboards derived from the index were available to the citizens only in their final, “ready-made” state, such as “what is the level of compliance to the lockdown in your home city compared to other cities”? discouraging reflection and encouraging the very thing they were supposedly measuring – compliance.
Lack of Ministry of Health Data
Many of the unclarities above could have been addressed – or at least mitigated – by clear and transparent reporting of Covid-19 infections and transmissions data by the governmental body responsible, Ministry of Health (MoH). Yet, despite repeated requests, throughout the first month of pandemic the data was published by the ministry’s spokesperson as cropped images on the ministry’s Telegram channel. This required a dedicated manual input of the information by volunteers to keep track of the official numbers. Even later, with a new updated data dashboard, users could not receive numerical information and moreover each new version overwrote the previous one. Data scientist Dan Bareket who kept those previous versions manually has shown that there are gross differences between those older and newer versions.
Those episodes come together to showcase how data can be wielded as a cudgel rather than a precise tool: collected through undisclosed means and used to create popular compliance, suppressing discussions of measures or alternatives.
About the authors
Alex Gekker is Assistant Professor in New Media and Digital Culture at the University of Amsterdam. He writes about how sociotechnical systems are designed to influence users, and his research touches upon maps and surveillance, quantification, the datafication of society, the experience economy, and interface critique. He has co-edited two Open Access books, Time for Mapping: Cartographic Temporalities (Manchester University Press, 2018) and Playful Mapping in the Digital Age (Institute of Network Cultures, 2016). In the past he has worked in a variety of media positions, as journalist, editor, and spokesperson.
Anat Ben-David is Senior Lecturer in the department of Sociology, Political Science and Communication at the Open University of Israel.