How to cite this article / Come citare questo articolo
Zito, L. (2026). Neurolabour: Information Asymmetry, Power Imbalance and the Governance of Cognitive Surveillance in Employment Relationships. Aequitas Magazine, 6, 1-13.
DOI: https://doi.org/10.5281/zenodo.20538710
ABSTRACT (EN) The proliferation of neurotechnology in occupational settings — encompassing electroencephalography (EEG) wearables, functional near-infrared spectroscopy (fNIRS) headsets, eye-tracking systems, and galvanic skin-response sensors — has inaugurated a new frontier in workplace governance that legal scholars, labour economists, and bioethicists have barely begun to map. This article interrogates the structural information and power asymmetries that neuro-monitoring technologies engender in employment relationships. Drawing on a critical analysis of agency theory, labour law doctrine, neuroimaging epistemology, and human rights frameworks, we argue that continuous neurophysiological surveillance transforms the employer-employee dyad in ways that are qualitatively distinct from prior forms of workplace monitoring. Employers gain real-time, granular, largely unverifiable access to workers’ cognitive and affective states — attention, mental workload, stress, emotional valence — while employees remain epistemically and legally exposed, lacking both interpretive parity and adequate regulatory protection. We propose a multi-layered normative framework premised on data minimisation, co-determination, neuro-specific consent, and algorithmic accountability, and we call for legislative intervention at national and supranational levels to forestall an emerging regime of cognitive enclosure.
ABSTRACT (IT) La proliferazione delle neurotecnologie nei contesti lavorativi — comprensiva di dispositivi EEG indossabili, sistemi fNIRS per la spettroscopia nel vicino infrarosso, tecnologie di eye-tracking e sensori di risposta galvanica cutanea — ha inaugurato una nuova frontiera nella governance del lavoro che giuristi, economisti del lavoro e bioeticisti hanno appena cominciato a esplorare. Il presente articolo esamina le asimmetrie strutturali di informazione e potere che le tecnologie di neuro-monitoraggio generano nei rapporti di lavoro subordinato. Attraverso un’analisi critica della teoria dell’agenzia, della dottrina giuslavoristica, dell’epistemologia del neuroimaging e dei framework dei diritti fondamentali, si sostiene che la sorveglianza neurofisiologica continua trasforma la relazione datore di lavoro-lavoratore in termini qualitativamente distinti rispetto alle forme tradizionali di controllo. Il datore di lavoro acquisisce un accesso in tempo reale, granulare e in larga misura inverificabile agli stati cognitivi e affettivi del lavoratore — attenzione, carico mentale, stress, valenza emotiva — mentre quest’ultimo rimane esposto sul piano epistemico e giuridico, privo tanto di parità interpretativa quanto di adeguata protezione normativa. Si propone un framework normativo a più livelli fondato sulla minimizzazione dei dati, sulla codeterminazione, sul consenso neuro-specifico e sulla responsabilità algoritmica, auspicando un intervento legislativo a livello nazionale e sovranazionale per prevenire l’affermarsi di un regime di enclosure cognitiva.
SUMMARY: 1. INTRODUCTION – 2. INFORMATION ASYMMETRY AND POWER IN EMPLOYMENT: THEORETICAL FOUNDATIONS – 3. THE EPISTEMIC PROPERTIES OF NEUROPHYSIOLOGICAL DATA – 4. THE LEGAL LANDSCAPE: GAPS AND INADEQUACIES – 5. TOWARDS A NORMATIVE FRAMEWORK FOR GOVERNING NEUROLABOUR – 6. OBJECTIONS AND RESPONSES – 7. CONCLUSION.
1. INTRODUCTION
In the spring of 2023, a logistics company operating across seven European countries announced that it would equip warehouse supervisors with EEG-enabled headbands capable of detecting fatigue and distraction in real time. The initiative was framed in the language of wellbeing: ostensibly, the data would trigger automated alerts encouraging workers to take breaks before errors occurred. Labour unions representing the affected workforce responded with immediate concern, identifying a fundamental asymmetry: management would possess continuous visibility into workers’ neurological states, while workers would have no reciprocal access to how that data was processed, stored, or used in performance evaluations. The episode encapsulates a tension at the heart of what we term neurolabour — the deployment of neurotechnology to measure, interpret, and act upon the cognitive and affective states of employees.
The architectonics of this asymmetry have long haunted the cultural imagination. In Fritz Lang’s Metropolis (1927), the city’s subterranean workforce toils in conditions of absolute epistemic subordination: the workers know only the machine they tend; the master of Metropolis, Joh Fredersen, surveys the entire productive apparatus from his glass tower above the clouds, commanding an informational vantage point that is itself the primary instrument of domination. Lang’s visual grammar — the panoptical control room, the mechanical rhythms imposed on human bodies, the radical invisibility of those who labour to those who govern — prefigures with remarkable precision the structural logic of neuro-monitoring in the contemporary workplace. What has changed is not the fundamental architecture of that dominion but its substrate: where Lang’s capitalist read the bodies of workers through spatial organisation and mechanical tempo, the twenty-first-century employer reads them through electroencephalographic signatures, cortisol proxies, and prefrontal oxygenation indices. The tower has become an algorithm; the machine-room has become the skull.
Nearly seven decades later, Andrew Niccol’s Gattaca (1997) transposed Lang’s structural intuition into a biological key. Where Metropolis rendered domination visible through stone and steel, Gattaca imagined a society in which involuntary biological data — a drop of blood, a strand of hair — determined a person’s professional destiny with algorithmic finality, unseen and unchallengeable. The film’s central conceit, that the body’s unguarded signals could be harvested to rank, sort, and exclude workers without their meaningful consent, was received as dystopian speculation. Nearly three decades after its release, that speculation has become operational. What Niccol dramatised through the lens of genetic determinism, contemporary neurotechnology is enacting through a different but structurally analogous substrate: the continuous, involuntary electromagnetic and haemodynamic activity of the working brain. Together, these two cinematic visions — Lang’s panoptical tower and Niccol’s genetic panopticon — delineate the twin axes of the problem this article addresses: the concentration of informational power in the hands of those who govern labour, and the reduction of the worker to a body whose innermost states are legible to others but opaque, in their institutional consequences, to herself.
Neurolabour is not a marginal phenomenon. Writing in the Harvard Business Review, Nita Farahany — whose monograph The Battle for Your Brain (St. Martin’s Press, 2023) constitutes the most comprehensive account to date of neurotechnology’s encroachment into everyday life — documented that tens of thousands of workers worldwide were already using neurophysiological monitoring devices in occupational settings as of 2023, with the global neurotech market expanding at a compound annual rate of approximately 12% and projected to reach USD 21 billion by 2026 (Farahany, 2023a; 2023b). Independent market analyses corroborate this trajectory: the global wearable EEG headsets market alone was valued at USD 122 million in 2023 and is forecast to reach USD 245 million by 2030 (Grand View Research, 2024), while broader wearable brain device revenues are projected to exceed USD 1.1 billion by 2030 (ResearchAndMarkets, 2025). Five thousand companies worldwide are reported to have deployed SmartCap fatigue-monitoring technology for workers in mining, aviation, and heavy industry (Farahany, 2023b; Pham et al., 2023). The technologies involved span a broad spectrum: passive EEG wristbands that infer attention from brainwave signatures; fNIRS headsets that measure oxygenation in the prefrontal cortex as a proxy for cognitive load; pupillometry systems embedded in workplace screens; and multimodal platforms that fuse heart-rate variability, skin conductance, and facial action unit recognition into composite ‘cognitive state’ dashboards. Each technology operates on a common logic: extracting signals from the body that are, in most circumstances, invisible to third-party observers, and rendering them legible, quantifiable, and actionable by employers.
The legal and ethical literature has begun to engage with these developments, but primarily through the lens of privacy — treating neuro-monitoring as an especially invasive form of the data collection that digital workplaces have long normalised. This framing, while important, is insufficient. The distinctive wrong at stake is not merely one of privacy violation but of structural epistemic and power asymmetry. When an employer can read a worker’s cognitive state with greater fidelity than the worker can articulate it, a new form of dominion is established — one that operates beneath the threshold of conscious negotiation and that current legal frameworks are poorly equipped to address.
This article proceeds as follows. Section 2 situates neuro-monitoring within the broader theoretical literature on information asymmetry and power in employment relationships. Section 3 analyses the specific epistemic properties of neurophysiological data and the asymmetries they generate. Section 4 surveys the current legal landscape at EU and international levels, identifying the gaps neuro-monitoring exposes. Section 5 develops a normative framework for governing neurolabour. Section 6 considers objections, and Section 7 concludes.
2. INFORMATION ASYMMETRY AND POWER IN EMPLOYMENT: THEORETICAL FOUNDATIONS
2.1 The Principal-Agent Problem and Its Limits
Orthodox principal-agent theory holds that the employment relationship is structured by the employer’s (principal’s) difficulty in observing and verifying the agent’s (employee’s) effort and disposition (Jensen & Meckling, 1976; Holmström, 1979). From this perspective, monitoring technologies are efficiency-enhancing: they reduce the information gap that generates moral hazard, enabling incentive contracts more tightly calibrated to actual performance. Neuro-monitoring, on this view, represents the logical culmination of a long trajectory from Taylorist time-and-motion studies to digital productivity tracking — a more precise instrument for resolving the observation problem.
This framing is, however, conceptually inverted in at least one critical respect. Principal-agent models presuppose that the observation problem runs primarily from principal to agent: it is the employer who cannot see what the employee is doing. Neuro-monitoring partially corrects this asymmetry — but it simultaneously creates an inverse asymmetry of equal or greater normative significance. Employees are, in most cases, unable to observe how their neurophysiological data are processed, what inferences are drawn from them, how those inferences feed into algorithmic management systems, or how the resulting classifications affect their working conditions, advancement prospects, or employment security. The information gap is not eliminated; it is relocated and, in important respects, deepened.
2.2 Power, Domination, and Epistemic Injustice
Beyond efficiency considerations, the employment relationship is a site of structural power imbalance. The foundational labour law insight — that the individual employee is in a position of structural dependency vis-à-vis the employer — is well established in comparative labour law scholarship (Freedland & Countouris, 2011; Davidov, 2016). What neuro-monitoring adds is a new vector of domination operating at the level of cognition. Philip Pettit’s republican conception of domination as the capacity for arbitrary interference (Pettit, 1997) is apt here: the employer who can monitor an employee’s stress response but is not legally required to disclose what is done with that information possesses a form of power over the employee that is, in Pettit’s terms, non-dominated — unchecked by accountability structures.
Miranda Fricker’s concept of epistemic injustice is equally germane (Fricker, 2007). Testimonial injustice occurs when a speaker’s credibility is systematically deflated due to identity prejudice; hermeneutical injustice occurs when a gap in collective interpretive resources disadvantages the less powerful party. Neuro-monitoring risks instantiating both. Workers who dispute algorithmically generated assessments of their cognitive state — insisting, for instance, that elevated cortisol readings reflect commuting stress rather than workplace disengagement — are structurally disadvantaged in that dispute: they lack the interpretive authority vested by technical expertise and the evidentiary standing conferred by access to the raw data. This is hermeneutical injustice enacted through biometric quantification.
2.3 Algorithmic Management and the Intensification of Asymmetry
Neuro-monitoring does not operate in isolation. It is most consequentially deployed as an input to algorithmic management systems — software platforms that aggregate real-time worker data and translate it into automated decisions about task allocation, performance scoring, shift scheduling, and disciplinary escalation (Rosenblat & Stark, 2016; Duggan et al., 2020). In this context, the information asymmetry is doubly compounded: not only does the employer possess neurophysiological data that the worker cannot see; the algorithmic translation of that data into consequential outcomes is itself opaque. Workers subject to such systems face what we might call a double epistemic wall: they cannot see the inputs, and they cannot see the logic connecting inputs to outputs.
3. THE EPISTEMIC PROPERTIES OF NEUROPHYSIOLOGICAL DATA
3.1 Granularity, Continuity, and Involuntariness
Neurophysiological signals differ from other forms of workplace data in three analytically important respects. First, they are characterised by extraordinary granularity. A wrist-worn EEG device sampling at 256 Hz generates over 900,000 data points per hour per channel. Even after artifact removal and feature extraction, the resulting dataset encodes information about cognitive state at a resolution that exceeds any self-report measure. This granularity enables inferences that go well beyond the proximate construct of interest: a classifier trained to predict ‘attention’ from EEG signals may incidentally encode information about fatigue, emotional valence, medication effects, neurological conditions, or susceptibility to persuasion.
Second, neurophysiological monitoring is continuous in a way that transforms the phenomenology of surveillance. Prior forms of workplace monitoring — CCTV, email scanning, keystroke logging — generate discontinuous records of behaviour. Neuro-monitoring generates a stream of data about inner states, potentially without interruption over entire working days or shifts. This continuity eliminates the cognitive and affective respite that discretised monitoring affords: there is no moment ‘off camera’ when the worker can recover, recalibrate, or simply be unobserved.
Third, and most fundamentally, neurophysiological signals are involuntary. A worker can choose what to type, what to say, where to look (to some degree), and how to comport themselves bodily. They cannot choose their cortical oscillations, pupillary dilation, or electrodermal response. This involuntariness has profound implications for consent: classical informational self-determination presupposes an agent who can choose what to reveal. Neuro-monitoring operates on a substrate that is, in significant measure, beyond the worker’s control — rendering consent frameworks both procedurally necessary and substantively insufficient as protective mechanisms.
3.2 The Interpretation Gap and Its Power-Laden Closure
Neurophysiological signals are not self-interpreting. The inferential chain from raw signal to attributed mental state is long, contested, and heavily mediated by theoretical commitments embedded in the software’s algorithmic architecture. The claim that elevated theta-band EEG power in frontal channels indicates ‘reduced attention’ rests on a substantial body of neuroscientific literature — but also on a set of assumptions about ecological validity, individual differences, and confound control that are actively disputed within that literature (Borghini et al., 2014; Zander & Kothe, 2011).
The employer who deploys a neuro-monitoring platform typically does not engage with this contestation. They receive a dashboard: a simplified representation — a score, a colour, a flag — that presents contested inference as established fact. The vendor’s algorithmic model functions as an epistemic black box that forecloses debate about interpretive validity precisely because its internal logic is proprietary. Workers who wish to challenge the inference — ‘my elevated theta power reflects deep concentration, not inattention’ — face not merely an uphill evidentiary battle but a structural incapacity to engage at the level at which the dispute must be joined.
3.3 Secondary Inferences and the Scope Creep of Cognitive Surveillance
A further dimension of the epistemic asymmetry concerns secondary inferences — information derived from neurophysiological data beyond the ostensible purpose of the monitoring programme. Machine learning models trained on EEG data have demonstrated the capacity to infer psychiatric diagnoses (depression, anxiety, ADHD), neurological conditions (early-stage Parkinson’s, epileptic tendency), personality traits, and susceptibility to coercion, with varying but commercially significant degrees of accuracy (Al-Ezzi et al., 2020; Rashid et al., 2019). An employer who deploys a ‘fatigue monitoring’ system receives not merely fatigue data but a rich neurological profile whose inferential potential extends far beyond the stated purpose. The worker, who consented (if they consented) to fatigue monitoring, has no visibility into the secondary inferences being drawn. This is a form of what Zuboff (2019) terms ‘behavioural surplus’ transposed into the neurological domain — what we term cognitive surplus extraction.
4. THE LEGAL LANDSCAPE: GAPS AND INADEQUACIES
4.1 The GDPR Framework and Its Limits
The General Data Protection Regulation (GDPR) constitutes the primary legal framework governing neuro-monitoring in European employment contexts. Neurophysiological data — EEG signals, heart rate variability, galvanic skin response — plainly qualify as personal data under Article 4(1) GDPR and, where they reveal health information, as special category data under Article 9. The processing of special category data in employment contexts requires, under Article 9(2)(b), authorisation by Member State law providing ‘appropriate and specific measures to safeguard the fundamental rights and the interests of the data subject’, with particular attention to Articles 88 GDPR’s requirements for employment-specific national implementing legislation.
The problem is that Article 88 is, in most Member States, underimplemented relative to the challenge that neuro-monitoring poses. National implementing provisions tend to address conventional forms of employment data processing — payroll, performance appraisals, email monitoring — without anticipating the continuous, involuntary, and inferentially expansive character of neurophysiological surveillance. Moreover, the lawful basis most frequently invoked by employers — contractual necessity under Article 6(1)(b) or legitimate interests under Article 6(1)(f) — is difficult to sustain in relation to neuro-monitoring. The Article 29 Working Party (now EDPB) has consistently held that consent in employment contexts is presumptively invalid due to the power imbalance between employer and employee (EDPB, 2020). Yet the EDPB’s guidance has not crystallised into binding determinations that specifically address neurophysiological monitoring.
4.2 The AI Act and Neuro-Monitoring
The EU Artificial Intelligence Act (Regulation 2024/1689), in force since August 2024, introduces a risk-based classification of AI systems relevant to neuro-monitoring. Systems that use biometric data — including physiological and behavioural signals — to infer sensitive attributes including health status are categorised, under Annex III, as high-risk AI systems when deployed in employment contexts. High-risk systems are subject to requirements of conformity assessment, technical documentation, human oversight, and transparency to affected persons.
However, the AI Act’s employment provisions (Article 26) focus on obligations of transparency to deployers (employers) rather than to affected persons (workers). Workers are entitled, under Article 50(1), to be notified when they interact with AI systems, but the obligation applies only where the AI output is ‘not obvious from the circumstances’ — a limitation that, as applied to background neuro-monitoring, is likely to be interpreted restrictively. More significantly, the AI Act does not address the interpretive asymmetry: it does not require employers to disclose the specific algorithmic logic by which neurophysiological signals are translated into performance assessments, nor to provide workers with meaningful contestation mechanisms calibrated to the cognitive-state inference context.
4.3 Emerging Neurorights Frameworks
At the constitutional and international law levels, a small but growing movement advocates for the recognition of ‘neurorights’ — fundamental rights specifically protective of cognitive liberty, mental privacy, mental integrity, and psychological continuity (Ienca & Andorno, 2017; Yuste et al., 2021). The most sustained and influential popular-scholarly case for such rights in the employment context has been made by Nita Farahany, whose monograph The Battle for Your Brain: Defending the Right to Think Freely in the Age of Neurotechnology (Farahany, 2023b) advances a comprehensive argument for ‘cognitive liberty’ as a foundational right encompassing the freedom to use, refuse, and maintain control over neurotechnology and the data it generates. Farahany documents the deployment of devices such as Emotiv’s MN8 enterprise earbuds — which monitor workers’ emotional and cognitive states through unobtrusive in-ear electrodes — and argues that the convergence of improving sensor fidelity, declining hardware costs, and algorithmic processing power is transforming what was recently a research-laboratory capability into an operational workplace management tool. A complementary line of inquiry, developed with particular rigour in the Italian constitutional scholarship, concerns the normative status of pre-expressive mental states — the cognitive, emotional, and deliberative processes that unfold prior to any externalisation of thought. Zito (2026) argues that the traditional notion of forum internum rests on an implicit technological premise — the physical inaccessibility of the mind — which advances in computational neuroscience and brain-reading technologies are progressively eroding, thereby generating a structural normative gap that existing constitutional frameworks are ill-equipped to address (Zito, 2026). This analysis is directly pertinent to the employment context: a neuro-monitoring device that captures pre-expressive cognitive states — the worker’s attentional drift before it issues in any observable behaviour, the stress response before it is vocalised or acted upon — penetrates precisely that domain which the forum internum doctrine was designed, however imperfectly, to protect. Chile became the first country to constitutionalise neurorights in 2021; Spain, France, and several Latin American countries have introduced or enacted related legislation. The Neurotechnology and Human Rights project at Columbia University has proposed a draft international protocol on neurorights.
These initiatives are significant but, in their current form, insufficiently attentive to the employment context. Constitutional neurorights frameworks tend to be framed at a high level of generality — prohibiting ‘arbitrary interference with brain data’ — without engaging with the structural power dynamics that make the employment deployment of such technologies distinctively problematic. The bioiuridical dimension of neurorights — the articulation of neurotechnological challenges in terms of fundamental rights grounded in the biological integrity of the person — has been explored in the Italian literature by Zito (2024), who situates the neurorights debate within a broader framework of equity and transparency as necessary conditions for any legitimate application of neurotechnology, and underscores the urgency of regulatory intervention before normalisation forecloses meaningful choice. The worker’s situation differs from the patient’s or the consumer’s: the employment relationship creates a background coercive context in which formal legal rights may be substantively hollow if workers face implicit or explicit adverse consequences for exercising them.
4.4 Collective Labour Law: An Underexplored Terrain
A significant gap in the literature concerns the role of collective labour law — in particular, works council rights, trade union information and consultation rights, and collective bargaining — in constraining neuro-monitoring. In jurisdictions with strong co-determination traditions (Germany, the Netherlands, the Nordic countries), works councils possess significant rights to be consulted or to co-determine the introduction of technical monitoring systems. The German Works Constitution Act (Betriebsverfassungsgesetz, §87(1)(6)) requires works council co-determination for the introduction or use of technical devices designed to monitor the behaviour or performance of employees — a provision that plainly extends to neuro-monitoring platforms.
However, collective rights are unevenly distributed across sectors and jurisdictions. In the gig economy, in non-unionised service sectors, and in countries with weak collective labour law traditions, workers face the information and power asymmetries of neuro-monitoring without any institutional counterweight. The result is a two-tier landscape in which the employees most likely to be subjected to intensive cognitive surveillance — warehouse workers, call centre agents, delivery drivers — are precisely those least likely to benefit from collective legal protections.
4.5 The Italian Perspective: Article 4 of the Workers’ Statute and Cognitive Surveillance
The Italian legal system offers a particularly instructive case study for the governance of neuro-monitoring, combining a strong constitutional tradition of labour protection with a legislative framework on remote surveillance that, while modernised by the Jobs Act reform of 2015, remains structurally untested against the challenge of neurophysiological data. Article 4 of the Workers’ Statute (Legge 20 maggio 1970, n. 300, as amended by D.Lgs. 14 settembre 2015, n. 151) constitutes the primary regulatory instrument governing remote monitoring of workers by technological means. Its current formulation, introduced to supersede a provision widely considered obsolete in the digital era, establishes that audiovisual equipment and other tools from which the possibility of remote monitoring of workers’ activity may derive can be employed exclusively for organisational and productive needs, workplace safety, or the protection of company assets (“tutela del patrimonio aziendale”), and only following the conclusion of a collective agreement with the unitary trade union representation (RSU) or the company trade union representations (RSA). Where no agreement is reached, authorisation may be sought from the territorially competent Labour Inspectorate (Ispettorato Nazionale del Lavoro, INL), which may impose conditions and limitations on the deployment.
The critical interpretive question, which Italian courts and the INL have not yet directly addressed in the neuro-monitoring context, concerns whether neurophysiological monitoring platforms fall within the scope of Article 4(1) — instruments “from which the possibility of remote monitoring of workers’ activity may derive” — or within the scope of Article 4(2), which excludes from the prior-authorisation requirement tools that are “used by the worker to perform their work” (strumenti di lavoro). The distinction is consequential: for Article 4(2) tools, the employer need not obtain trade union agreement or INL authorisation, and the data may be used for all purposes connected to the employment relationship, subject only to an obligation to provide adequate information to the worker. The temptation for employers to characterise neuro-monitoring wearables as “work tools” — safety devices or productivity aids rather than surveillance instruments — is evident and must be resisted by systematic interpretation. The Italian Supreme Court (Corte di Cassazione) has consistently held, in relation to analogous technologies, that the qualification of a device as a “work tool” under Article 4(2) must be assessed on the basis of its primary function, not its formal designation by the employer: where the primary or dominant function of a device is control of the worker, Article 4(1) applies regardless of any ancillary productive utility (Cass. n. 15391/2024; Cass. n. 25732/2021).
Applied to neuro-monitoring, this jurisprudential approach yields the following analysis. A wearable EEG device deployed to monitor workers’ attention, stress, or cognitive load does not perform any function in the productive process other than generating data about the worker’s neurophysiological state. Even where the ostensible purpose is safety — alerting the worker to fatigue before error occurs — the device’s primary functional output is a continuous stream of data about the worker’s inner cognitive condition, accessible by and transmitted to the employer. This plainly falls within the scope of Article 4(1): prior trade union agreement or INL authorisation is required. The data collected, furthermore, may be used “for all purposes connected to the employment relationship” under Article 4(3) only where adequate prior information has been provided to the worker — a requirement that, in the context of neurophysiological monitoring, must be interpreted to encompass disclosure of the algorithmic logic, the inferences drawn, and the uses to which those inferences may be put in employment decisions.
Article 4 also intersects with the Italian implementation of GDPR through the Codice della Privacy (D.Lgs. 30 giugno 2003, n. 196, as amended by D.Lgs. 10 agosto 2018, n. 101). The Garante per la protezione dei dati personali has not yet issued specific guidance on neurophysiological workplace monitoring, but its established position on biometric data processing in employment contexts — requiring a specific legal basis, strict proportionality assessment, and prohibition of use for purposes beyond those specified — is directly applicable. The Garante’s provvedimenti on videosorveglianza and email monitoring confirm that Italian data protection supervision is responsive to emerging technological challenges, and a request for specific guidance on neuro-monitoring would be both legally appropriate and strategically advisable for employers contemplating deployment. What Article 4, read in conjunction with the Codice della Privacy and GDPR, unambiguously establishes is that neuro-monitoring without prior trade union agreement or INL authorisation, and without provision of adequate information to workers about the nature and uses of the monitoring, is unlawful under Italian law — exposing the employer to administrative sanctions under Article 38 of the Workers’ Statute and to criminal liability under Article 171 of the Codice della Privacy.
5. TOWARDS A NORMATIVE FRAMEWORK FOR GOVERNING NEUROLABOUR
5.1 Principles of the Framework
We propose a normative framework for the governance of neuro-monitoring in employment organised around five core principles: (1) cognitive liberty as a fundamental labour right; (2) strong data minimisation; (3) interpretive co-determination; (4) neuro-specific informed consent and its structural limits; and (5) algorithmic accountability and contestation.
Cognitive liberty — the right to mental self-determination, understood as encompassing both the freedom to use and the freedom to refuse neurotechnology — should be recognised as a fundamental labour right, operationalised within employment law as a right not to be subjected to neurophysiological monitoring as a condition of employment, a condition of continued employment, or a condition of access to benefits or advancement. This operationalisation would require Member States to enact legislation, under Article 88 GDPR, that specifically prohibits the use of neuro-monitoring data in individual performance assessment or disciplinary proceedings, absent exceptional justification and works council co-determination.
5.2 Data Minimisation and Purpose Limitation
The principle of data minimisation, already embedded in Article 5(1)(c) GDPR, must be operationalised with neuro-specific force. This requires, at minimum, a presumption against continuous monitoring: collection should be triggered only upon specific safety-relevant events (identified in advance by reference to validated thresholds), and aggregated at population rather than individual level wherever the putative purpose can be served without individual attribution. Purpose limitation must include a prohibition on secondary inference: vendors and employers should be legally prohibited from deriving inferences from neurophysiological data beyond those specifically authorised at the point of collection.
5.3 Interpretive Co-Determination and Symmetry
The interpretive asymmetry that characterises neuro-monitoring — employers receive algorithmic outputs; workers receive nothing — should be addressed through a right of interpretive co-determination. This principle would require, in practice, that any classification generated by a neuro-monitoring system used in employment decisions be made available to the affected worker in intelligible form; that the worker be entitled to contest that classification before any adverse employment consequence materialises; that the contestation mechanism be capable of engaging with the validity of the underlying neuroscientific inference and not merely with procedural compliance; and that works councils or trade unions be entitled to commission independent technical audits of neuro-monitoring algorithms.
5.4 Neuro-Specific Consent and Its Limits
Consent, as noted, is structurally inadequate as the primary regulatory mechanism for neuro-monitoring in employment. Nonetheless, a reformed consent regime has an important ancillary role. Consent to neuro-monitoring should be: specific (articulating the precise signals collected, the algorithms applied, and the inferences authorised); layered (distinguishing between collection, analysis, and use in employment decisions); revocable without adverse consequence; and accompanied by a right to receive the raw data and the algorithmic output upon request. Most importantly, consent should not be available as a lawful basis for the use of neuro-monitoring data in individual disciplinary or termination decisions: such uses should require, at minimum, works council co-determination and, where they involve special category data, explicit Member State legislative authorisation.
5.5 Algorithmic Accountability and Transparency
Transparency of algorithmic decision-making — already partially addressed by GDPR’s Article 22 (automated decision-making) and the AI Act’s high-risk system requirements — must be extended and deepened in the neuro-monitoring context. We propose, specifically, a right of workers and their representatives to access the training data composition, validation methodology, and accuracy characteristics of any neuro-monitoring algorithm that generates outputs used in employment decisions; a requirement that vendors publish algorithmic audits by accredited independent bodies; a duty of vendors to disclose material accuracy limitations, including known differential performance across demographic groups; and a prohibition on the use of algorithms that cannot demonstrate ecological validity in the specific occupational setting in which they are deployed.
6. OBJECTIONS AND RESPONSES
6.1 The Welfare Argument
The most common objection to restrictive neuro-monitoring regulation is the welfare argument: if neurotechnology can identify fatigue before errors occur, reduce workplace accidents, or enable personalised job design that prevents burnout, prohibition or heavy restriction sacrifices worker welfare in the name of abstract rights. This objection has genuine force in specific safety-critical contexts — aviation, surgical environments, nuclear plant operation — where early detection of cognitive degradation may prevent catastrophic harm.
The response is not blanket prohibition but contextual calibration. Safety-critical uses of neuro-monitoring, where the benefit is concrete and the alternative is greater risk of physical harm, can be accommodated within our framework through a heightened-scrutiny exception subject to proportionality analysis, mandatory works council co-determination, and strict prohibition on the use of safety data for performance management. The welfare argument cannot, however, justify the wholesale deployment of neuro-monitoring for productivity optimization, performance ranking, or attentional fine-tuning in ordinary commercial contexts — the uses that are, empirically, the most common.
6.2 The Innovation Objection
A related objection holds that restrictive regulation will chill the development of beneficial neurotechnology and disadvantage European employers in global markets. This objection conflates productive and extractive forms of neuroinnovation. The development of neurotechnology to treat neurological disease, to support communication for people with ALS, or to enable environmental design for cognitive accessibility is not impeded by labour law protections that restrict the use of neuro-monitoring in employment decisions. What is restricted is the capacity to extract cognitive surplus from workers without accountability — an activity whose productive value is contested and whose distributional consequences are regressive.
6.3 The Equivalence Objection
A third objection holds that neuro-monitoring is not qualitatively different from other forms of workplace surveillance — CCTV, call recording, keystroke logging — that are already accepted with appropriate safeguards. If we accept monitoring of behaviour, why not monitoring of the cognitive processes that produce behaviour? The objection underestimates the significance of the involuntariness, the inferential expansiveness, and the interpretive asymmetry identified in Section 3. Behaviour is, in principle, within the agent’s control; cognition is not. A worker can choose what to say in a recorded call; they cannot choose their cortisol levels. Neuro-monitoring does not merely extend behavioural surveillance to a new domain; it reconfigures the ontological boundary between the self and the observable world of employment.
7. CONCLUSION
Neurolabour — the deployment of neurotechnology in employment contexts — represents a qualitative transformation in the structure of information and power in the employment relationship. The technologies at issue are not merely more efficient versions of conventional monitoring; they operate on a substrate of involuntary, inferentially expansive, and interpretively asymmetric data that existing legal frameworks are inadequately equipped to address.
The framework we propose — organised around cognitive liberty as a fundamental labour right, strong data minimisation, interpretive co-determination, reformed consent doctrine, and algorithmic accountability — seeks to restore a degree of epistemic symmetry to the employment relationship without prohibiting the genuinely beneficial uses of neurotechnology in occupational contexts. Implementation will require legislative action at both EU and Member State levels, active engagement by data protection authorities and labour inspectorates, and strengthened collective bargaining rights for workers in sectors where neuro-monitoring deployment is most intensive.
The broader stakes extend beyond the employment relationship. How societies govern cognitive surveillance in the workplace will prefigure how they govern it in other domains — education, health, public space — and will shape the emerging norms of what it means to be a human subject in a world of proliferating neurotechnology. The choices made now, while the technologies are nascent and the regulatory frameworks malleable, will prove extraordinarily difficult to reverse once cognitive enclosure has been normalised. The argument for intervention is not merely one of individual rights but of the structural conditions of democratic citizenship: a workforce subjected to pervasive, asymmetric, and legally unaccountable neurological surveillance is a workforce whose capacity for the autonomous political and social agency that democracy requires has been systematically diminished.
REFERENCES
Al-Ezzi, A., Kamel, N., Faye, I., & Gunaseli, E. (2020). Review of EEG, ERP, and brain connectivity estimators as predictive biomarkers of social anxiety disorder. Frontiers in Psychology, 11, 730. https://doi.org/10.3389/fpsyg.2020.00730.
Borghini, G., Astolfi, L., Vecchiato, G., Mattia, D., & Babiloni, F. (2014). Measuring neurophysiological signals in aircraft pilots and car drivers for the assessment of mental workload, fatigue and drowsiness. Neuroscience & Biobehavioral Reviews, 44, 58-75.
Davidov, G. (2016). A Purposive Approach to Labour Law. Oxford University Press.
Duggan, J., Sherman, U., Carbery, R., & McDonnell, A. (2020). Algorithmic management and app-work in the gig economy: A research agenda for employment relations and HRM. Human Resource Management Journal, 30(1), 114-132.
Farahany, N. A. (2023a). Neurotech at Work: Welcome to the World of Brain Monitoring for Employees. Harvard Business Review, 101(2), 43–48.
Farahany, N. A. (2023b). The Battle for Your Brain: Defending the Right to Think Freely in the Age of Neurotechnology. St. Martin’s Press.
Freedland, M., & Countouris, N. (2011). The Legal Construction of Personal Work Relations. Oxford University Press.
Fricker, M. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford University Press.
Holmström, B. (1979). Moral hazard and observability. Bell Journal of Economics, 10(1), 74-91.
Ienca, M., & Andorno, R. (2017). Towards new human rights in the neurotechnology age. Life Sciences, Society and Policy, 13(1), 1-27.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360.
Pettit, P. (1997). Republicanism: A Theory of Freedom and Government. Oxford University Press.
Pham, T., Tran, D., Ma, W., & Tran, S. N. (2023). Neurosurveillance in the workplace: do employers have the right to monitor employees’ minds? Frontiers in Human Dynamics, 5, 1245619. https://doi.org/10.3389/fhumd.2023.1245619.
Rashid, M., Sulaiman, N., Mustafa, M., Khatun, S., & Bari, B. S. (2019). The classification of EEG signal using different machine learning techniques for BCI application. In J.-H. Kim, H. Myung, & S.-M. Lee (Eds.), Robot Intelligence Technology and Applications. RiTA 2018. Communications in Computer and Information Science (Vol. 1015, pp. 207–221). Springer. https://doi.org/10.1007/978-981-13-7780-8_17.
Rosenblat, A., & Stark, L. (2016). Algorithmic labor and information asymmetries: A case study of Uber’s drivers. International Journal of Communication, 10, 3758-3784.
Yuste, R., Goering, S., Arcas, B. A. Y., Bi, G., Carmena, J. M., Carter, A., … & Wolpaw, J. (2021). Four ethical priorities for neurotechnologies and AI. Nature, 551(7679), 159-163.
Zander, T. O., & Kothe, C. (2011). Towards passive brain-computer interfaces: Applying brain-computer interface technology to human-machine systems in general. Journal of Neural Engineering, 8(2), 025005. https://doi.org/10.1088/1741-2560/8/2/025005.
Zito, L. (2024). Neurotecnologie: cenni sulla dimensione biogiuridica dei “neurodiritti”. Aequitas Magazine, 1–9. https://doi.org/10.5281/zenodo.17888295.
Zito, L. (2026). Oltre il forum internum: libertà cognitiva e stati mentali pre-espressivi. Aequitas Magazine, 5, 1–19. https://doi.org/10.5281/zenodo.20023026.
Zoli, C. (2016). Il controllo a distanza dell’attività dei lavoratori e la nuova struttura dell’art. 4, legge n. 300/1970. Diritto del Lavoro e Relazioni Industriali.
Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. PublicAffairs.
