ClearFactAll topics

Should governments use AI-powered surveillance systems during pandemics to monitor disease spread?

5 peer-reviewed studies for, 5 against

Key takeaways

Stronger evidence: Against

Verdict

The evidence slightly weighs against government use of AI-powered surveillance systems during pandemics, though the calculus is finely balanced. The theoretical case for AI surveillance is strong: modeling in Science demonstrated that digital contact tracing can suppress epidemic spread faster than manual methods, and Nature Medicine documented genuine public health benefits from digital monitoring during COVID-19. However, real-world deployment told a different story. Most contact tracing apps globally failed to achieve the adoption thresholds needed for epidemiological effectiveness — typically estimated at 60% or more of the population. Meanwhile, the privacy costs were substantial and persistent. Research in Nature Biotechnology documented that surveillance apps routinely collected data far exceeding public health needs, and governments have a consistent historical pattern of converting emergency surveillance powers into permanent infrastructure. Israel repurposed its domestic intelligence agency for pandemic contact tracing. China's health code system became embedded in daily life well beyond COVID. South Korea's detailed exposure notifications inadvertently revealed individuals' movements, relationships, and identities. The fundamental problem is asymmetry: surveillance infrastructure is easy to build during emergencies when civil liberties objections are overridden, but nearly impossible to dismantle afterward. Narrowly scoped, privacy-preserving approaches like decentralized Bluetooth proximity detection represent a reasonable middle ground, but the broader question of AI-powered government surveillance during pandemics introduces risks to civil liberties that real-world evidence shows are consistently realized rather than merely theoretical.

Strongest study on each side

Ranked by citation count.

Strongest study for

Digital technologies in the public-health response to COVID-19 (Nature Medicine, 2020)

Jobie Budd, Benjamin S. Miller et al. · 886 citations

Review demonstrated that AI-powered digital contact tracing, symptom surveillance, and mobility tracking significantly accelerated outbreak detection and response times during COVID-19, enabling targeted interventions that reduced transmission while minimizing the need for broad lockdowns.

Strongest study against

Apps Gone Rogue: Maintaining Personal Privacy in an Epidemic (Arxiv, 2020)

150 citations

Argued that pandemic surveillance apps collect far more data than necessary for disease tracking, lack meaningful consent frameworks during emergencies, and that governments have a documented history of repurposing crisis surveillance infrastructure for permanent monitoring programs long after emergencies end.

Open the interactive page →

The evidence

For (5)

Digital technologies in the public-health response to COVID-19 (Nature Medicine, 2020)

Jobie Budd, Benjamin S. Miller et al. · 886 citations

Review demonstrated that AI-powered digital contact tracing, symptom surveillance, and mobility tracking significantly accelerated outbreak detection and response times during COVID-19, enabling targeted interventions that reduced transmission while minimizing the need for broad lockdowns.

Systematic Determinants of Global COVID-19 Burden: Longitudinal Time-Series Analysis Using Big Data-Driven Artificial Intelligence (Journal of Medical Internet Research, 2025)

Zicheng Cao, Wenjie Han et al.

A retrospective longitudinal time-series study across 38 countries analyzing AI-driven machine learning approaches to understand how viral variants, population immunity, environmental factors, and public health measures collectively determined COVID-19 burden. The research demonstrates AI's effectiveness in integrating multifaceted surveillance data to quantify disease transmission and severity patterns over time.

Against (5)

Apps Gone Rogue: Maintaining Personal Privacy in an Epidemic (Arxiv, 2020)

150 citations

Argued that pandemic surveillance apps collect far more data than necessary for disease tracking, lack meaningful consent frameworks during emergencies, and that governments have a documented history of repurposing crisis surveillance infrastructure for permanent monitoring programs long after emergencies end.

Investigating Citizens' Acceptance of Contact Tracing Apps: Quantitative Study of the Role of Trust and Privacy (JMIR mHealth and uHealth, 2024)

Grace Fox, Lisa van der Werff et al.

Study examining factors fostering trust and privacy perceptions in contact tracing apps using structural equation modeling at two time points. Findings show that trust and perceived privacy significantly influence citizens' willingness to adopt surveillance apps and disclose personal data, highlighting barriers to effective pandemic surveillance adoption.

Privacy and Surveillance Attitudes During Health Crises: Acceptance of Surveillance and Privacy Protection Behaviours (Technology in Society, 2021)

Athina Ioannou and Iis Tussyadiah

This study examines public attitudes toward AI surveillance during pandemics, revealing that people are deeply concerned about data collection and monitoring of online activities despite willingness to accept some surveillance measures. It demonstrates the significant psychological and behavioral costs of surveillance systems, with privacy protection behaviors becoming necessary responses to perceived threats.

The debate

Victor: COVID-19 killed nearly seven million people. The Science modeling showed that digital contact tracing could have suppressed transmission faster than any manual system. South Korea and Taiwan used AI surveillance to identify and isolate cases within hours — not days — and they kept death tolls a fraction of countries that relied on traditional methods. When a pandemic is killing thousands daily, AI surveillance is not optional. It is a moral obligation.

Andre: South Korea's system also publicly broadcast people's movements so precisely that neighbors could identify who visited which bars, who was having affairs, and who was in a mental health clinic. People stopped getting tested because they feared the exposure notifications more than the virus. That is not a theoretical privacy concern — that is documented harm that directly undermined the public health goal.

Victor: Those are implementation failures, not fundamental flaws. Decentralized systems like the Apple-Google Exposure Notification framework proved you can do proximity detection without location tracking. The technology exists to build privacy-preserving surveillance. The question is whether governments should use AI-powered tools during pandemics — and the answer is clearly yes, with appropriate safeguards.

Andre: Appropriate safeguards that have never actually materialized in practice. Nature Biotechnology documented that pandemic apps routinely collected far more data than needed. Israel used Shin Bet — its domestic intelligence agency — for contact tracing. China's health code system is now a permanent fixture of daily life. The pattern is consistent across every government, democratic or authoritarian: emergency surveillance powers expand and never contract. You are not arguing for a temporary tool. You are arguing for permanent infrastructure.

Victor: Then mandate sunset clauses, independent oversight, and data deletion requirements. We require these for wiretaps and FISA warrants. The framework for constraining surveillance within democracies exists. And the alternative — relying solely on manual contact tracing during a fast-moving respiratory pandemic — is to accept thousands of preventable deaths while protecting an abstract principle.

Andre: It is not abstract. The right to move freely without government tracking is foundational to democratic life. And the practical evidence does not even support your case. Most contact tracing apps worldwide failed to reach the 60% adoption threshold needed for epidemiological effectiveness. The UK app, the French app, the Australian app — all fell far short. You are trading fundamental rights for a tool that demonstrably did not work at scale.

Victor: Adoption was low partly because of exactly the kind of privacy panic you are promoting. If governments invested in trustworthy, transparent, privacy-preserving systems with genuine public engagement, adoption would follow. Singapore's TraceTogether eventually reached 80% penetration. The technology works when implemented with public trust.

Andre: Singapore is a city-state of six million people with a social contract fundamentally different from most democracies. And TraceTogether's data was accessed by police for criminal investigations despite explicit government promises it would only be used for contact tracing. That is my entire argument in one case study: promises made during emergencies do not survive contact with institutional incentives. Every surveillance system ever built has been repurposed. Every one.

Victor: And every pandemic without adequate surveillance tools has killed millions. I will take the imperfect tool with guardrails over no tool at all.

Andre: And I will take the slower, more manual approach that does not hand governments a surveillance architecture they will never give back.

Frequently asked questions

Did contact tracing apps work during COVID-19?

Modeling suggested app-based tracing could suppress spread if enough people used it. In practice, a review of deployed apps found most failed to reach the adoption rates needed for epidemiological effectiveness.

What are the privacy risks of pandemic surveillance apps?

Research found apps often collected far more data than disease tracking required and lacked meaningful consent during emergencies. Some created location and social-connection databases that remained after the crisis.

Why do people hesitate to use contact tracing apps?

Studies show trust and perceived privacy strongly influence whether people adopt these apps and share personal data. Many people report concern about data collection and monitoring, even when they accept some surveillance.

Is there a privacy-friendly way to do digital contact tracing?

Narrowly scoped, privacy-preserving approaches such as decentralized Bluetooth proximity detection are seen as a reasonable middle ground. Reviews also stress that AI tools work best when combined with human expertise.

Related questions