Private Algorithms Supersede Democracy
Denver's Flock Safety Wrongful Accusation
Futurism reported that Denver, Colorado, relied on Flock Safety data, which led to a wrongful accusation. An idle Android phone transmits approximately 900 data points to Google in 24 hours. Campaign Zero and the University of Michigan's Center for Local, State, and Urban Policy (CLOSUP) found that Motorola Solutions holds ALPR contracts with over 60 police departments; in Michigan, 26% of sheriffs and police chiefs use ALPRs, and 10% use facial recognition technology. The NYPD and LAPD utilize Clearview AI's database and Palantir's Gotham system, Campaign Zero reports. Frontiers in Political Science illustrates how centralized digital platforms, such as India's Aadhaar biometric identity system and Singapore's Smart Nation program, can de-democratize local institutions internationally. Hongfei Gu in Journal of Chinese Political Science and the Observatory for Information and Democracy document that the French company Civipol has developed biometric identity systems for Senegal and Côte d’Ivoire. This "surveillance federalism" redefines citizens as data subjects rather than political participants, Frontiers in Political Science argues.
Detailed 'Data Doubles' Limit Agency
Policy Review adds that the creation of detailed "data doubles" further limits agency by enabling social sorting, manipulation, and discrimination. Platforms "herd" billions of users into artificially curated spaces where algorithmic interpretation, rather than user intent, steers information exposure, Swati Srivastava explains. Swati Srivastava, in Perspectives on Politics, states that "systems are designed to 'pre-empt agency, spontaneity, and risk' by using granular behavioral data to orchestrate human situations and foreclose alternative actions." This continuous direction of user interactions toward greater engagement diminishes the ability to make truly autonomous decisions, Policy Review details.
Proprietary Algorithms Shielded by Trade Secrets
The Knight First Amendment Institute at Columbia University and Ezenwaohaetorc Journals contend that systemic accountability is fundamentally challenged because proprietary algorithms are shielded by expanded trade secret laws and intellectual property rights. The technical complexity of machine learning models makes them difficult for even experts to audit, limiting the insight needed for meaningful public or regulatory review, the Knight First Amendment Institute at Columbia University explains. The Knight First Amendment Institute at Columbia University also notes that traditional regulatory mechanisms, such as the Freedom of Information Act (FOIA), do not extend to the private sector, and recent legal decisions have empowered agencies to withhold confidential information. Specific proprietary algorithms, such as Facebook's DeepFace facial recognition software, its News Feed and "Trending Topics" recommendation systems, Amazon's recruitment algorithm, and Google's search and news feed algorithms, all remain shielded by trade secrets, according to Swati Srivastava, the Knight First Amendment Institute at Columbia University, and Ezenwaohaetorc Journals.
DMA and DSA Do Not Dismantle Asymmetry
Swati Srivastava and the Knight First Amendment Institute at Columbia University point out that while the EU's Digital Markets Act (DMA) and Digital Services Act (DSA) include provisions for mandatory data sharing with trusted researchers, these frameworks do not dismantle the asymmetric information control that allows powerful platforms to function as quasi-sovereign entities. Scholarship from Columbia Law School's Blue Sky Blog and the Cornell Law School Scholarship Repository, alongside analyses by ACT Online, CCIA, RePEc, and Emilia Smolak Lozano in Frontiers in Political Science, demonstrates that these regulations have compelled tech firms to undertake structural redesigns rather than superficial adjustments, indicating that the "Brussels Effect" can lead to substantive compliance outcomes. These measures often act as reactive attempts to constrain power after it has already crystallized, primarily accommodating corporate interests rather than fundamentally reasserting democratic control, Ezenwaohaetorc Journals argues.
Big Tech Operates as 'Data Sovereigns'
Hongfei Gu, in Journal of Chinese Political Science, states that this perpetuates an asymmetric power dynamic where Big Tech operates as 'data sovereigns' beyond effective public oversight. Swati Srivastava and Policy Review predict that the continuous algorithmic steering and creation of "data doubles" will further diminish the capacity for autonomous choice, leading to a future where personal decisions are increasingly pre-empted and manipulated by opaque systems. The Knight First Amendment Institute at Columbia University and Ezenwaohaetorc Journals observe that core algorithmic governance will remain insulated from democratic review due to proprietary algorithms shielded by trade secrets and technical complexity. Without a fundamental reconstruction of legal authority, technical might will continue to dictate legal right, further eroding public trust and the deliberative foundations necessary for democracy, Stanford University warns.
Comments ()