Surveillance is an asymmetry of seeing before it is a technology, and monitoring is where a graduate meets it

Surveillance names a relation before it names a device. One party observes, another is observed, and the second cannot usually see how, how long, or to what end. This seventh of ten units argues that the ethical problem sits in that unequal relation rather than in the sensor. The consequences are administrative: an unnamed lawful basis, a retention period nobody chose and a reuse nobody disclosed matter more than the camera or the keyboard logger. A graduate meets the relation in the workplace, where the standard the law sets becomes checkable.

Surveillance is a relation of unequal seeing

Marx (2002) attacks the dictionary definition of surveillance as close observation of a suspected person. He shows that much surveillance is applied categorically, to places, spaces, networks and categories of person rather than to suspects, and from the technology, the collection process and the data he draws twenty-eight dimensions for describing a means of surveillance. Those dimensions refuse the intuition that a new sensor is a new ethics: what varies is the scope and target, and the relation stays unequal. Ball (2010) establishes that organisations and surveillance go hand in hand and that workplace surveillance takes both social and technological forms. Her survey places monitoring among longstanding managerial practices rather than recent technology.

The economics changed, and the relation of seeing did not

Zuboff (2015) states the change in economic terms. In the logic she calls surveillance capitalism, data extraction and new contractual forms follow from computer-mediated transactions. Marx's (2002) dimensions do the same work, separating a security camera from an activity-scoring tool without treating either as unprecedented.

The Federal Trade Commission's case against Facebook gives the arrangement a document. The complaint recorded that the company told members it would not share their personal information with advertisers and did. It also overrode privacy settings without notice. The settlement required express consent before any change that overrode those settings, with independent privacy audits every two years for twenty years (FTC 2011). The binding instrument was a consent order between a regulator and the party that held the record.

Monitoring is where a graduate meets the asymmetry

The Information Commissioner's Office (2022) publishes the guidance a UK graduate is most likely to meet, requiring a lawful basis and a documented retention schedule, stating that workers' expectations of privacy are higher at home than in the workplace, and holding that a customer making monitoring a condition of business does not justify it. Rosenblat and Stark (2016) supply the empirical case. In a nine-month study of Uber driver experiences they found the platform exercises significant indirect control over how drivers do their jobs, and concluded that the information and power asymmetries produced by the application are fundamental to that control.

The failures that matter are administrative, not technical

Moore (2000) separates the moral question from the legal one, the method this unit follows. He argues that knowledge of the kinds of surveillance used at a company should be explicit to employees, and that some evaluative monitoring violates privacy rights and should not be used in most cases, so disclosure is a condition of legitimate evaluation. Ajunwa, Crawford and Schultz (2017) trace employer surveillance from the Pinkerton detectives of the 1850s to the apps that quantify productivity. They argue that data analytics and mobile device design let productivity scoring fall outside the protections built for earlier forms.

The outcome evidence has arrived, and the governance question remains

Ball (2021), in a report for the Publications Office of the European Union, re-evaluates the literature on monitoring in the standard workplace and in home working during the COVID-19 pandemic, mapping practice across European jurisdictions.

Ravid, White and colleagues (2023) meta-analysed 94 independent samples covering 23,461 workers. They found no evidence that monitoring improves performance, together with raised stress that did not depend on the characteristics of the scheme. The legal argument therefore no longer runs ahead of a missing literature. It runs ahead of one that reports the observing relation as the harm.

Directive (EU) 2024/2831, which member states must transpose by 2 December 2026, requires decisions that restrict, suspend or terminate an account to be taken by a human being, gives the worker a right to an oral or written explanation of any automated decision, requires oversight staff with the training and authority to override, and bans monitoring of emotional state, private conversations and conduct outside working time (European Parliament and Council 2024). The last two bans name conditions the cited schemes leave out: they fall hardest on the worker whose health, care burden or second job the scheme makes visible.

Contested question: is the harm the collection, the use, or the relation?

The literature disagrees about where the wrong sits. The collection position holds that the harm is the intrusion into personal information and the loss of control over disclosure. Moore (2000) occupies that ground in reading evaluative monitoring as a violation of privacy rights, and Bhave, Teo and Dalal (2019) describe it when they centre employees' expectation of control. The use position holds that the wrong is what the data does once it exists. Ravid and colleagues (2019) treat monitoring as a performance instrument whose effects on the worker are the question. The relation position, this unit's, holds that the asymmetry itself is the harm, so collection and use are symptoms. Zuboff (2015, 2019) makes the relation the object of analysis and Marx (2002) supplies the categorical reach. What would settle it is outcome evidence tying a practice to a harm while holding other features constant, so the three readings give different predictions. Ravid, White and colleagues (2023) supply part of that test, since stress rose regardless of the scheme's characteristics, which the relation position predicts, though the measure is self-reported.

The three readings do not rest on equal evidence. The collection position is a normative argument from Moore (2000) and a review of expectations in Bhave, Teo and Dalal (2019); the use position is a further review in Ravid and colleagues (2019); the relation position is theory from Zuboff (2015, 2019) and Marx (2002) with one meta-analysis in Ravid, White and colleagues (2023), whose outcomes are self-reported. Only the last supplies outcome data, and it measures what the worker says. The counter-case section carries the objection to this unit's thesis. While the dispute is open, an employer should name the lawful basis, set and keep a retention period, and be able to explain the scheme to the worker it observes.

The counter-case: the asymmetry buys coordination

The strongest objection is that the observing relation is a condition of organised work rather than an ethical defect, so treating it as the harm over-convicts. Ball (2010) is the anchor, since a firm that monitored nothing could not coordinate the work it exists to do, and the Information Commissioner's Office (2022) supports the milder version, monitoring being lawful once a basis and a schedule are in place. The reply is that the objection concerns how much a scheme matters and how it is governed, not whether the relation exists.

The human who operates the system absorbs the accountability

Alkhatib and Bernstein (2019) argue that algorithms must bridge the gap between high-level policy and on-the-ground decisions. They make inferences in novel situations the policy or training data do not cover, so the human who acts on the output absorbs the accountability for it. The system's contribution hides behind a colleague's decision. Kellogg and colleagues (2020) name the six mechanisms by which algorithmic control operates, so the routes through which an employer acts on a worker become visible rather than diffuse.

Where the argument is heading

The direction the cited work sets is a turn from describing the asymmetry to building instruments that can test it. Both reviews close with a research agenda whose next task is measuring outcomes rather than stating principles, and Ravid, White and colleagues (2023) hold the characteristics of a scheme constant and find stress that does not move with them, which is the first measured result of that kind. The eighth unit asks what an audit can show.

The sources this unit rests on

Each note is this unit's summary of what the cited document did and found, not a quotation from it.

• Marx, G. T. (2002). What's New About the "New Surveillance"? Classifying for Change and Continuity. Surveillance & Society. doi:10.24908/ss.v1i1.3391 An article that rejects the dictionary definition of surveillance as observation of a suspect and shows much surveillance is applied categorically to places, networks and categories of person. Its twenty-eight dimensions, drawn from the technology, the collection process and the data, separate a new sensor from a new ethics.
• Ball, K. (2010). Workplace surveillance: an overview. doi:10.1080/00236561003654776 A review establishing that organisations and surveillance go hand in hand and that workplace surveillance takes both social and technological forms. It is the evidence that monitoring is a longstanding managerial practice rather than a recent technology.
• Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. PublicAffairs. https://www.publicaffairsbooks.com/titles/shoshana-zuboff/the-age-of-surveillance-capitalism/9781610395694/ The book that carried the asymmetry argument into public argument, examining a form of power it calls surveillance capitalism and the corporate quest to predict and control behaviour. It names the relation between the party holding the data and the party described.
• Zuboff, S. (2015). Big Other: Surveillance Capitalism and the Prospects of an Information Civilization. Journal of Information Technology. doi:10.1057/jit.2015.5 The article stating the asymmetry in economic terms. It reads data extraction and analysis and new contractual forms as consequences of computer-mediated transactions, making the data the product rather than a by-product.
• Information Commissioner's Office (2022). Employment practices and data protection: monitoring workers. https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/employment/monitoring-workers/ The guidance a UK graduate is most likely to meet, distinguishing what an employer must, should and could do and requiring a lawful basis and a documented retention schedule. It states that workers' expectations of privacy are higher at home than in the workplace and that customer pressure does not justify monitoring.
• Rosenblat, A. and Stark, L. (2016). Algorithmic labour and information asymmetries: a case study of Uber's drivers. International Journal of Communication. https://ijoc.org/index.php/ijoc/article/view/4892 The empirical case where the worker cannot see the system that evaluates them. A nine-month study of Uber driver experiences found the platform exercises significant indirect control, and that the information and power asymmetries it produces are fundamental to that control.
• Moore, A. D. (2000). Employee Monitoring and Computer Technology: Evaluative Surveillance V. Privacy. Business Ethics Quarterly. doi:10.2307/3857899 An article separating the moral question from the legal one. It argues that knowledge of the kinds of surveillance used at a company should be made explicit to employees, and that some evaluative monitoring violates privacy rights and should not be used in most cases.
• Ajunwa, I., Crawford, K. and Schultz, J. M. (2017). Limitless worker surveillance. California Law Review. doi:10.15779/Z38BR8MF94 A law review tracing employer surveillance from the Pinkerton detectives of the 1850s to the apps that quantify productivity today. It argues data analytics and mobile device design let productivity scoring fall outside the protections built for older monitoring.
• European Parliament and Council (2024). Directive (EU) 2024/2831 on improving working conditions in platform work. https://eur-lex.europa.eu/eli/dir/2024/2831/oj Requires transposition by 2 December 2026. It bans monitoring of emotional state, private conversations and conduct outside working time, requires decisions that restrict, suspend or terminate an account to be taken by a human being, and gives the worker a right to a written explanation of an automated decision.
• Ball, K. (2021). Electronic monitoring and surveillance in the workplace. Publications Office of the European Union. doi:10.2760/5137 A policy report re-evaluating the literature on monitoring in the standard workplace and in home working during the COVID-19 pandemic and mapping practice across European jurisdictions. Its scope is legal and policy output, the unit's own limit stated in miniature.
• Bhave, D. P., Teo, L. H. and Dalal, R. S. (2019). Privacy at Work: A Review and a Research Agenda for a Contested Terrain. Journal of Management. doi:10.1177/0149206319878254 A review framing employee privacy as a contest of expectations. It records that employees expect control of the personal information and access they give while employers expect extensive information and access, organising the conflict without resolving it.
• Ravid, D. M., Tomczak, D. L., White, J. C. and Behrend, T. S. (2019). EPM 20/20: A Review, Framework, and Research Agenda for Electronic Performance Monitoring. Journal of Management. doi:10.1177/0149206319869435 An earlier synthesis of the electronic performance monitoring literature. Observing no comprehensive review since 2000, it finds technology now lets organisations monitor to a greater extent and intensity than before, and reports contradictory findings on the effects.
• Ravid, D. M., White, J. C., Tomczak, D. L., Miles, A. F. and Behrend, T. S. (2023). A meta-analysis of the effects of electronic performance monitoring on work outcomes. Personnel Psychology. doi:10.1111/peps.12514 A meta-analysis of 94 independent samples covering 23,461 workers. It found no evidence that monitoring improves performance and an association between monitoring and increased worker stress that did not depend on the characteristics of the scheme, while more transparent and less invasive monitoring drew more positive worker attitudes. It is the measured outcome evidence the earlier framing said was missing.
• Kellogg, K. C., Valentine, M. A. and Christin, A. (2020). Algorithms at Work: The New Contested Terrain of Control. Academy of Management Annals. doi:10.5465/annals.2018.0174 A review organising what algorithmic control does to work. Using Edwards's contested terrain, it finds algorithmic control operates through six mechanisms it names, supplying the vocabulary the unit's settlement instrument would need to vary.
• Alkhatib, A. and Bernstein, M. (2019). Street-Level Algorithms. Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems. doi:10.1145/3290605.3300760 A paper applying street-level bureaucracy theory to algorithms that must bridge high-level policy and on-the-ground decisions. The human who acts on the output absorbs the accountability for it, hiding the system's contribution behind a colleague's decision.
• Federal Trade Commission (2011). Facebook settles FTC charges that it deceived consumers by failing to keep privacy promises. https://www.ftc.gov/news-events/news/press-releases/2011/11/facebook-settles-ftc-charges-it-deceived-consumers-failing-keep-privacy-promises Records charges that Facebook promised members it would not share their personal information with advertisers and then shared it, and the order requiring express consent before overriding privacy settings plus independent privacy audits for twenty years. It dates the observing relation in a regulatory document.