The Environmental Costs of Surveillance Capitalism: A Case Study of Social Media Platforms

Technical Summary: The Environmental Costs of Surveillance Capitalism

Problem Statement
The paper addresses a significant gap in the literature regarding the environmental impact of "surveillance capitalism," a business model predicated on the extraction and commodification of behavioral data for profit. While the social harms of this model are well-documented, its material environmental costs remain largely unquantified. The authors argue that the infrastructure required to sustain surveillance capitalism—including telemetry, user tracking, algorithmic engagement optimization, and large-scale data storage—contributes substantially to the rising carbon footprint of Information and Communication Technologies (ICT). Current environmental assessments often fail to distinguish between resources used for core platform functionality and those consumed specifically by surveillance and profit-maximization processes.

Methodology
To quantify these costs, the authors propose a conceptual framework linking surveillance processes to material infrastructure and validate it through a comparative case study.

1. Conceptual Framework: The authors define three categories of computational resource usage:

   • Core & Navigation Features: Sufficient functionalities required for basic platform use.
   • User Generated Data & Content: Data created by users.
   • Surveillance & Tracking: Resources consumed by data extraction, analysis, and targeted advertising.
     The framework distinguishes between "sufficient" functionalities and "overhead" functionalities (excess resource consumption driven by corporate incentives).

2. Case Study Design: The study compares two functionally equivalent microblogging platforms with divergent business models:

   • X (formerly Twitter): A corporate social media (CSM) platform driven by profit, user tracking, and algorithmic feeds.
   • Mastodon: A decentralized, non-commercial alternative (Alternative Social Media or ASM) based on open protocols and lacking corporate incentives for engagement maximization.

3. Experimental Setup:

   • User Journeys: The authors developed a bespoke measurement tool using Playwright (for browser automation), MitmProxy (for intercepting HTTPS traffic), and a SQLite database.
   • Scenarios: Three user journeys were simulated and repeated 100 times each: (1) Scrolling the main feed for ~5 minutes, (2) Posting a message and viewing it, and (3) Scrolling a specific content feed (TechCrunch) for ~5 minutes.
   • Metrics: The system captured HTTP requests/responses, URLs, cookies, total data transfer sizes, and timestamps.
   • Quantification: Network traffic data was converted to energy usage (kWh) using coefficients from Istrate et al. (0.0591 kWh/GB) and then to carbon emissions (gCO2e) using the global grid intensity factor (445 gCO2e/kWh). The results were scaled to X's reported daily active minutes (8 billion) and daily tweet volume (500 million) to estimate annual emissions.

Key Contributions

1. Conceptual Framework: The paper introduces a framework that maps the processes of data capture and commodification to their material realities, serving as a research agenda for quantifying the environmental costs of surveillance capitalism.
2. New Metrics: The authors define two specific metrics:
   • Corporate Overhead: The proportion of network traffic in a CSM exceeding the sufficiency baseline established by an ASM.
   • User Tracking Overhead: A finer-grained metric isolating traffic specifically attributable to user data collection (e.g., specific API endpoints and cookies).
3. Empirical Baseline: The study provides an initial empirical investigation into digital sufficiency by demonstrating how non-commercial alternatives can serve as a baseline for minimum necessary resource consumption.

Results
The comparative analysis of network traffic yielded the following findings:

• Corporate Overhead:

  • In the main feed scenario (User Journey 1), corporate overhead accounted for 86.04% of X's data traffic.
  • In the posting scenario (User Journey 2), it accounted for 49.78%.
  • In the TechCrunch feed scenario (User Journey 3), it accounted for 55.15%.
  • Emission Estimates: Based on these proportions and X's scale, the authors estimate a lower bound for annual CO2e emissions attributable to corporate overhead.
    • Using User Journey 2 and 3 data: ~123.88 ktCO2e per year.
    • Using User Journey 1 (assuming it represents typical usage): ~1.47 MtCO2e per year.

• User Tracking Overhead:

  • This specific category (traffic directly linked to tracking endpoints like /i/api/1.1/graphql/user_flow.json) was significantly smaller, accounting for 1.66% to 6.65% of total traffic depending on the journey.
  • Emission Estimates: The lower bound for annual emissions from user tracking overhead is approximately 14.89 ktCO2e.

Significance and Claims
The paper modestly claims to establish a lower bound for the environmental impact of surveillance capitalism, acknowledging that network traffic is an imperfect proxy for total carbon emissions (as it excludes server-side processing, storage, and manufacturing impacts).

The authors emphasize several key points:

• Knowledge Asymmetry: Corporate social media platforms actively obfuscate their data extraction practices, making transparency difficult. The study highlights that even with technical expertise, researchers can only see the "tip of the iceberg" of these environmental costs.
• Digital Sufficiency: The results suggest that a significant portion of resource consumption in platforms like X is "corporate overhead" that does not contribute to the user experience but rather to profit generation and engagement maximization.
• Policy Implications: While user tracking overhead has a lower carbon footprint than general corporate overhead, it may be a more viable target for policy intervention as it does not directly challenge the core business model of content delivery. However, the authors argue that addressing the broader "corporate overhead" is necessary for genuine digital sufficiency.
• Research Agenda: The framework is presented as a tool for future research to investigate other elements of the infrastructure (e.g., server-side processing, user device UI) and other platform pairings (e.g., YouTube vs. PeerTube).

The paper concludes that quantifying these costs is essential for reducing the knowledge asymmetry that allows surveillance capitalism to operate with little resistance, thereby advancing both sustainability research and environmental justice.