Protecting Workers' Right to Mental Health in Digital Work Environments: Legal Challenges and Reform Directions in Vietnam

Introduction

A. Research problem and significance

Digital transformation is not only changing the tools of work but also restructuring managerial power within employment relationships. Email, messaging applications, timekeeping software, screen-monitoring tools, performance evaluation systems, and work-allocation algorithms enable enterprises to coordinate labour more rapidly, but at the same time increase their capacity to intervene in workers’ time, behaviour, and condition. An article on the OSHwiki of the European Agency for Safety and Health at Work (EU-OSHA) identifies digitalisation as a potential source of work intensification, reduced job control and autonomy, social isolation, and the perception or reality of being under surveillance.¹ The urgency of the issue is reflected in the scale of its impact. WHO estimates that 16% of working-age adults had a mental disorder in 2023, and that depression and anxiety result in around 12 billion working days lost each year, at a cost of US$1 trillion per year in lost productivity.²

In Vietnam, a cross-sectional study of 1,228 workers at two steel and ceramic tile manufacturing facilities recorded an overall prevalence of signs of mental health disorders of 20.8%; workers who self-reported stressful work were 1.35 times more likely to show such signs than those who did not perceive their work as stressful.³ Another study of 111 staff members and lecturers at Phenikaa University found prevalence rates of 42.3% for insomnia, 46.8% for anxiety, 21.6% for depression, and 26.1% for stress. Because of the limited sample sizes, these findings are not representative of the entire workforce, but they indicate that the issue is not confined to direct production work and is also significant among knowledge workers who use technology frequently.⁴

These figures do not justify equating all psychological problems with occupational diseases, nor do they establish technology as the sole cause. They do, however, demonstrate that mental health at work is an occupational risk-prevention issue, rather than merely a voluntary employee-benefit choice for enterprises or a matter of individual self-care. The core legal question is: in a work environment organised through data and continuous connectivity, how should the law allocate preventive obligations among the State, employers, technology providers, and workers?

B. Literature gap, objectives, and research questions

Existing scholarship has developed three relevant but insufficiently integrated lines of inquiry. Public-health and occupational-safety research identifies psychosocial hazards and recommends organisational interventions. Labour-law scholarship examines working time, the right to disconnect, and employer duties of care. A third body of work addresses privacy, workplace surveillance, and algorithmic management. Each line supplies an important part of the problem, but a privacy-only analysis cannot determine whether lawful monitoring nevertheless creates excessive work intensity, while a working-time analysis alone cannot address opaque scoring, automated discipline, or inferred mental-health data. These strands have not yet been adequately consolidated in the Vietnamese legal context into a single preventive framework linking work design, digital managerial power, mental health, and enforceable remedies.

The article therefore has three objectives: first, to conceptualise the right to mental health protection in digital work as an application of existing rights to health, safe working conditions, rest, privacy, equality, and participation; second, to test whether Vietnamese law supplies a coherent cycle of identification, assessment, control, consultation, and remedy for digital psychosocial risks; and third, to formulate reforms that are legally enforceable yet adaptable to enterprise size, job functions, emergencies, and new technologies.

It answers three research questions. First, which features of digitally organised work create legally cognisable psychosocial risks? Second, to what extent do international standards and selected comparative models translate those risks into preventive employer obligations and worker rights? Third, which amendments to Vietnamese labour and occupational safety and health law would close the identified gaps without treating all after-hours contact, monitoring, or algorithmic assistance as inherently unlawful?

C. Methodology, scope, and contribution

The article employs doctrinal analysis to interpret the 2013 Constitution, the 2019 Labour Code, the 2015 Law on Occupational Safety and Health, the 2025 Law on Personal Data Protection, and relevant implementing instruments. It synthesises interdisciplinary evidence to identify pathways through which work organisation may produce psychological harm. Functional comparison is then used to examine international standards and selected approaches in the European Union, France, and Australia. These jurisdictions are not presented as transplant models; they are selected because they represent complementary regulatory techniques: risk prevention, collective bargaining, and a qualified statutory right to refuse unreasonable after-hours contact.

The analysis concerns work-related psychosocial risks associated with digital communication, monitoring, algorithmic management, remote work, and platform work. It does not attempt to diagnose mental disorders or claim that every mental-health condition is occupationally caused. The article’s principal contribution is an integrated legal model that separates three questions often conflated in current debate: whether data processing is lawful, whether digital work design creates an unacceptable psychosocial risk, and which institution must prevent or remedy that risk.

The remainder of the article proceeds as follows. Part II develops the theoretical and legal foundations of mental-health protection in digital work. Part III examines international standards and comparative approaches. Part IV evaluates Vietnamese law and identifies regulatory gaps. Part V proposes a sequenced reform model, and Part VI concludes.

Theoretical and legal foundations

A. Mental health and psychosocial risks

WHO defines mental health as a state of mental well-being that enables people to cope with the stresses of life, realise their abilities, learn and work well, and contribute to their community; mental health has intrinsic and instrumental value and is a basic human right.⁵

In employment relationships, the “right to mental health protection” should not be understood as a right to require employers to guarantee that workers remain in a positive psychological state at all times. A reasonable conception of this right comprises three levels of obligation. First, the duty not to cause harm: work must not be organised, monitored, harassed, or data processed in a manner that creates a serious and foreseeable risk to mental health. Second, the preventive duty: psychosocial risks arising from work design must be identified, assessed, and controlled. Third, the duty of support and rehabilitation: workers experiencing mental health problems should be enabled to access assistance, reasonable accommodation, and return-to-work measures.

WHO lists workplace psychosocial risks as including excessive workload and work pace, long or inflexible working hours, lack of control, organisational cultures that enable negative behaviour, violence and harassment, discrimination, unclear roles, job insecurity, and work–family conflict.⁶

The concept of psychosocial risk has an important legal advantage: it shifts the focus from diagnosing an individual worker’s illness to the design and management of work. A worker may not yet have a mental disorder but may still be exposed to a risk that requires control. Conversely, not every psychological difficulty experienced by a worker can automatically be attributed to the enterprise. Responsibility arises only where a work-related risk is identifiable, capable of causing harm, and within the employer’s authority or capacity to control.

B. Mental health risks in digital work environments

First, continuous connectivity causes working time to “spill over” into rest periods. An after-hours message may take only a few minutes to answer, but the obligation to monitor a device continually makes psychological recovery difficult. The issue is not only the total amount of demonstrable overtime, but also a state of constant availability and the fragmentation of rest time.

Eurofound defines the right to disconnect as the right of workers to be able to disengage from work and refrain from engaging in work-related electronic communications during non-work hours. Its survey indicates that headaches, fatigue, and anxiety or stress are more common among those required to work additional hours, while health issues are reported by a smaller share of workers in companies that implement a right to disconnect.⁷

Second, digital monitoring changes workers’ perception of autonomy. The American Psychological Association’s 2023 Work in America survey found that 45% of workers monitored through technology reported that their workplace had a negative impact on their mental health, compared with 29% among workers who were not monitored.⁸ Technology may serve security, performance assessment, and worker-support purposes; however, prolonged recording of keystrokes, screenshots, location, voice, facial expressions, or task-completion speed may create a sense of constant observation. Risks increase where data-collection criteria are not transparent, data are reused for ranking or disciplinary purposes, and workers lack channels through which to obtain explanations or lodge complaints.

Third, algorithmic management may increase work intensity, uncertainty, and information asymmetry. Bowdler et al. synthesise evidence showing that algorithmic management is associated with intensified workloads, irregular and unpredictable working hours with unpaid waiting time, reduced decision authority, and social isolation; transparency and worker involvement are important mitigating factors.⁹

Fourth, violence and harassment may occur beyond the enterprise’s physical premises. Insulting messages, publicly ranking workers to humiliate them, disseminating images, or excluding a member from online work groups may all cause psychological harm. Digital traces may facilitate proof, but the speed of dissemination, round-the-clock accessibility, and blurred boundaries between personal accounts and work channels can deepen the harm.

Fifth, risks are distributed unevenly. Women workers often shoulder unpaid care responsibilities; platform workers face income uncertainty and algorithmic evaluation; persons with psychosocial disabilities are at risk of stigma; and middle managers may both experience pressure from the system and be required to transmit that pressure to subordinates. Accordingly, formally neutral policies may still produce differentiated adverse impacts.

Participatory research with women garment workers in Bangladesh, Sri Lanka and Vietnam, built around women workers’ own accounts, highlights the value of gender-responsive healthcare, social dialogue, decent and fair working conditions, and quality childcare for their mental health and well-being.¹⁰

C. Intersecting rights to health, rest, privacy, and equality

Article 12 of the International Covenant on Economic, Social and Cultural Rights recognises everyone’s right to the highest attainable standard of physical and mental health; Article 7 requires safe and healthy working conditions and rest, leisure and the reasonable limitation of working hours.¹¹ ILO Convention No. 155 requires the establishment of a coherent national policy on occupational safety, occupational health, and the working environment, with the aim of preventing accidents and injury to health arising out of, linked with, or occurring in the course of work. Vietnam ratified the Convention on 3 October 1994.¹²

Under these standards, the right to mental health protection in the digital environment need not be constructed entirely as a new right. It is a contemporary manifestation of already recognised rights: the right to safe working conditions, the right to rest, the right to privacy, the right to non-discrimination, and the right to participate in decisions affecting oneself. Nevertheless, identifying it separately has operational value because it clarifies the relevant risks, preventive obligations, and enforcement mechanisms in a context where technology blurs traditional boundaries.

International standards and comparative experience

A. WHO and ILO risk-management approaches

WHO’s 2022 Guidelines on mental health at work recommend organisational-level interventions to assess and modify working conditions; training for managers; improved mental health literacy among workers; reasonable accommodation for persons with mental health conditions; and return-to-work programmes.¹³ The joint WHO and ILO policy brief emphasises that measures to control risks should follow the hierarchy of control, giving priority to eliminating hazards at source. Measures that increase individual resilience, such as stress-management training, should be delivered only as part of a broader programme and are unlikely to be effective on their own; they cannot substitute for changes to workload, work schedules, communication practices, or mechanisms for addressing violence and harassment.¹⁴

This approach is significant for law-making: enterprise obligations should not stop at providing psychological counselling hotlines after harm has already occurred. The appropriate sequence is risk identification, worker consultation, risk assessment, control at source, effectiveness monitoring, and individual support. This is the familiar preventive logic of occupational safety and health law, extended from physical, chemical, and biological hazards to psychosocial risks.

ILO Convention No. 190 extends the concept of the “world of work” to work-related communications, including those enabled by information and communication technologies. Although Vietnam has not ratified the Convention, this standard still provides a directly relevant model for addressing harassment in digital environments.¹⁵

B. European Union and France: disconnection and social dialogue

The European Parliament’s Resolution of 21 January 2021 describes the right to disconnect as the right of workers not to engage in work-related activities or communications by means of digital tools outside working time, without facing adverse consequences, and calls on the Commission to propose a Union directive on minimum standards and conditions for its exercise.¹⁶ French law does not impose a single technical model. Article L. 2242-17 of the Labour Code includes the practical arrangements for exercising the right to disconnect and regulating the use of digital tools within the annual bargaining on gender equality at work and on quality of life and working conditions; where no agreement is reached, the employer must draw up a charter after consulting the Social and Economic Committee.¹⁷ Eurofound case studies show that legal recognition of the right to disconnect often encourages the adoption of collective agreements and company policies; however, effectiveness depends on organisational culture, workload, managerial behaviour, and monitoring mechanisms, rather than merely on shutting down email servers.¹⁸

The French experience suggests two lessons for Vietnam. First, the law should establish a minimum right and a non-retaliation principle, while implementation should be specified through workplace dialogue to accommodate shift work, time zones, on-call positions, and emergencies. Second, the right to disconnect is not an absolute right prohibiting all after-hours communication; its purpose is to enable workers to refuse unreasonable requests and to ensure that time actually spent performing work is recorded and remunerated.

C. Australia: a qualified right to refuse after-hours contact

From 26 August 2024 for most businesses and from 26 August 2025 for small businesses, Australian law permits employees to refuse to monitor, read, or respond to contact outside their working hours unless the refusal is unreasonable. Contact may come from the employer or, where it relates to the employee’s work, from third parties such as clients or members of the public.¹⁹ In assessing reasonableness, the Australian system considers the reason for the contact; how the contact is made and the degree of disruption; whether the worker is paid or otherwise compensated for remaining available outside working hours; the nature of the role and level of responsibility; and personal circumstances, including family or caring responsibilities. Disputes are first addressed at the workplace and may be brought before the Fair Work Commission.²⁰

The Australian model has the advantage of framing the right to disconnect as a conditional right of refusal, avoiding the mechanical interpretation that every after-hours message is unlawful. This is an appropriate technique for Vietnam to consider: legislation may specify assessment factors, while collective agreements, internal labour regulations, and contracts may establish notification procedures, lists of on-call positions, compensation levels, and emergency communication channels.

Vietnamese law and regulatory gaps

A. Existing legal foundations

The 2013 Constitution recognises the inviolability of private life in Article 21, the right to health protection and care in Article 38, and the right to work and to choose an occupation in Article 35. These rights provide a constitutional foundation for interpreting obligations to protect mental health in employment relationships.²¹

The 2019 Labour Code provides that workers have the right to work under conditions ensuring occupational safety and health; employers are obliged to comply with labour law, respect workers’ honour and dignity, and establish mechanisms for workplace dialogue.²²

Articles 105, 107, 109, and 111 of the Labour Code impose limits on normal working hours, overtime, rest breaks during working time, and weekly rest. In principle, time actually spent by workers performing required tasks outside normal hours should be identified and treated under the overtime regime.²³

The 2015 Law on Occupational Safety and Health requires employers to organise the assessment and control of dangerous and harmful factors; prepare occupational safety and health plans; and arrange health examinations and examinations for the detection of occupational diseases for workers.²⁴

Circular No. 19/2016/TT-BYT establishes requirements for occupational hygiene management, working-environment monitoring, and worker health management. The forms and training materials accompanying the Circular refer to the assessment of neuropsychological workload and of psychophysiological and ergonomic factors at work.²⁵

For sexual harassment, Decree No. 145/2020/ND-CP defines the “workplace” as including work-related locations and spaces, including communications conducted by electronic means. This is an important development because conduct occurring through email, messaging, or online platforms may therefore fall within the scope of regulation.²⁶

The 2025 Law on Personal Data Protection, effective from 1 January 2026, establishes principles concerning purpose, scope, and necessity, data-subject rights, and data-processing impact assessments; it also contains provisions on data protection in the recruitment, management, and use of employees and on data processing through artificial intelligence.²⁷ This framework is relevant to health data, biometric data, and behavioural data obtained through workplace monitoring.²⁸

B. Regulatory gaps

i. Psychosocial risks are neither defined nor systematically managed

Clause 5 Article 3 of the Law on Occupational Safety and Health defines a “harmful factor” as a factor that causes illness or impairs health during the course of work. As a matter of interpretation, this broad concept may encompass psychosocial factors. However, the Law does not separately define psychosocial risks; systems for monitoring, occupational hygiene records, and occupational diseases still operate primarily around physical, chemical, and biological agents and technically measurable workload.²⁹

A practical limitation is that enterprises find it difficult to determine the minimum compliance standard: must work pressure be surveyed, when must burnout risk be assessed, who is qualified to carry out the assessment, and what findings require work to be redesigned? Labour inspectors likewise lack criteria for determining when algorithmic target allocation, screen monitoring, or continuous online scheduling creates an unacceptable level of risk. In the absence of a mandatory procedure, the issue is often recognised only after a worker resigns, a conflict arises, or a clinical diagnosis is made.

This assessment is consistent with the position the ILO took at a May 2026 event in Hanoi marking Vietnam’s National Occupational Safety and Health Month: that psychosocial risks such as long working hours, chronic stress, violence, and harassment are shaped by how work is designed and managed and can therefore be prevented, and that Conventions Nos. 155 and 187, both ratified by Vietnam, can help strengthen the prevention and management of psychosocial hazards.³⁰

ii. Working-time rules do not address constant connectivity

Articles 105 and 107 of the Labour Code regulate normal working hours and overtime but do not establish a right to refuse after-hours contact or principles for determining working time where tasks are assigned through digital tools. Clause 2 Article 118 on internal labour regulations requires rules on working and rest time, but does not require enterprises to establish communication windows, on-call positions, response deadlines, or mechanisms for recording after-hours work.³¹

In practice, a series of small requests sent by message may not be recorded as overtime; workers may nevertheless feel compelled to monitor their devices for fear of being perceived as uncooperative. The “consent to overtime” mechanism under Article 107 is difficult to make meaningful where consent is not recorded for each occasion and the pressure to respond stems from organisational culture. As a result, rest time may exist on the timesheet while no longer being psychologically intact. This is a direct gap in current regulation and cannot be remedied simply by advising workers to turn off notifications themselves.

iii. Digital monitoring and algorithmic management remain outside working-condition regulation

The 2025 Law on Personal Data Protection creates an important layer of protection for data collection and processing. However, personal-data compliance and the protection of mental health are distinct tests. A monitoring system may have a lawful purpose, have been duly notified, and ensure data security, yet still cause harm if monitoring is excessive, work pace is intensified, or adverse decisions are generated that workers cannot understand or challenge.

The Labour Code does not require consultation before highly intrusive monitoring systems are deployed; it does not grant a right to explanation where algorithms determine shifts, scores, or disciplinary recommendations; and it does not require a simultaneous impact assessment of privacy, equality, and mental health. Disputes may therefore be reduced to the question of whether the enterprise “is permitted to collect the data”, while the substantive issue is how technology has restructured working conditions.

A 2026 ILO working paper on AI systems at work highlights emerging risks including intrusive surveillance, reduced autonomy, work intensification, and job insecurity, while also noting that existing occupational safety and health instruments, built on the premise that risks can be addressed at the design stage, do not account for risks that emerge dynamically as AI systems are deployed.³²

iv. Health and return-to-work mechanisms remain physically oriented

Article 21 of the Law on Occupational Safety and Health provides for periodic health examinations and examinations for the detection of occupational diseases, but does not require mental health risk assessment based on the nature of the work; does not establish specific confidentiality principles for psychological screening results; and does not impose an obligation to develop return-to-work programmes for workers absent due to work-related depression, anxiety, or burnout.³³

If an enterprise organises screening without adequate safeguards, psychological data may become a source of stigma or be used to remove workers from their positions. If no screening is conducted and no support channel exists, signs of risk may instead be overlooked. The legal gap therefore encompasses both the absence of a duty to provide support and the absence of safeguards for those receiving support. The law should distinguish clearly between roles: enterprises manage working conditions; health professionals conduct professional assessments; and workers are entitled to voluntariness, confidentiality, and freedom from adverse treatment for seeking assistance.

v. Protection has not kept pace with platform and remote work

Many platform workers are classified as partners or service providers and therefore have difficulty obtaining full access to rights relating to working time, occupational safety and health, and collective representation mechanisms. Yet this group is heavily managed through ratings, dynamic pricing, account deactivation, and automated work allocation. The Law on Occupational Safety and Health contains certain policies for persons working without employment contracts, but it does not impose psychosocial risk-management duties on platform enterprises that exercise algorithmic control.³⁴

For remote work performed under an employment contract, responsibility for ensuring safety in spaces controlled by the worker should be reasonably limited. However, enterprises still control workload, meeting schedules, monitoring tools, and response expectations. Current law does not clearly distinguish risks arising from the worker’s home environment from those generated by digital work design, creating a tendency to shift most responsibility onto the worker.

Recommendations for Vietnamese law reform

A. Recognise mental health protection and define psychosocial risks

First, point (b) of clause 1 Article 5 of the 2019 Labour Code should be amended to recognise that workers have the right to “work under conditions ensuring occupational safety and health; and to have their physical health, mental health, and privacy respected and protected in the course of work”. This addition would not make employers guarantors of health outcomes, but would establish a coherent interpretive basis for preventive obligations.

Second, Article 3 of the Law on Occupational Safety and Health should be supplemented with a definition: “Psychosocial risk means the possibility of harm to physical or mental health arising from the design, organisation, and management of work, social relations at the workplace, or the use of technology in the course of work.” At the same time, clause 5 Article 3 should be amended to confirm that harmful factors include psychosocial factors. This approach would integrate new risks into the existing preventive system without creating an entirely separate regime.

B. Require psychosocial-risk assessment and control

Article 18 of the Law on Occupational Safety and Health should be amended to require that the assessment and control of dangerous and harmful factors include psychosocial risks where warning signs exist or in sectors, occupations, or forms of work organisation identified as high-risk. A new clause should be added to Article 18 with minimum assessment elements covering: workload and work pace; working time and degree of after-hours connectivity; autonomy; violence and harassment; digital monitoring and algorithmic management; job insecurity; work–family conflict; and differentiated impacts on vulnerable groups.

The implementing decree should establish a six-step process: identification; consultation; assessment; prioritising elimination of risk at source; action planning; and reassessment. Small enterprises should be permitted to use simplified tools issued by the competent authority; enterprises using AI systems to manage labour or employing at least a specified number of workers should be required to prepare a written assessment report. Collective-level assessment results should be disclosed to workers’ representative organisations, without disclosing individual health data.

C. Introduce a qualified right to disconnect

A new Article 109a entitled “Right to disconnect” should be inserted after Article 109 of the Labour Code. Clause 1 should provide: “Workers have the right not to monitor, read, or respond to work-related communications outside working hours, unless the refusal is unreasonable under law, a collective labour agreement, or another lawful agreement.” Clause 2 should prohibit disciplinary action, discrimination, or adverse evaluation solely because a worker exercises this right.

Clause 3 of Article 109a should list criteria for assessing reasonableness: urgency and consequences of non-response; nature of the position and level of responsibility; notified on-call schedules; degree of disruption; frequency of contact; pay or compensation for remaining available; personal circumstances; and the availability of alternative channels. Clause 4 should allow internal labour regulations or collective agreements to prescribe communication windows, on-call positions, emergency channels, response times, and methods for recording actual working time.

At the same time, point (a) of clause 2 Article 118 of the Labour Code should be amended so that internal labour regulations must include “principles governing the use of digital communication tools, the exercise of the right to disconnect, and the recording of after-hours working time”. Article 107 should also be amended so that time actually spent handling work requests outside normal hours is aggregated and paid as overtime; consent to overtime should not be presumed merely because a worker has read a message.

D. Regulate monitoring and algorithmic management as labour-law issues

A new Article 118a on “electronic monitoring and algorithmic labour-management systems” should be added to the Labour Code. The new provision should require employers to provide advance notice of the types of data collected, purposes, retention periods, persons with access, the system’s principal logic, and consequences for workers; and to consult workers’ representative organisations before deploying technology capable of continuous monitoring or generating decisions that significantly affect recruitment, shift allocation, pay, performance assessment, discipline, or termination.

Article 118a should also establish limits: prohibit covert monitoring except where permitted by law; prohibit the use of emotion data or inferences about mental health to make adverse decisions without scientific and legal justification; guarantee the right to request human review of automated decisions; and require impact assessments addressing mental health, privacy, and equality. These obligations would supplement, not replace, obligations under the 2025 Law on Personal Data Protection.

E. Improve support and return-to-work arrangements

Clauses 1 and 2 Article 21 of the Law on Occupational Safety and Health should be amended so that periodic health examinations may include mental health screening appropriate to occupational risks, based on the principles of voluntariness and confidentiality, with the involvement of suitably qualified health personnel, and with results not to be used as a basis for discrimination. Mass screening should not be made compulsory, nor should human-resources personnel be tasked with diagnosing psychological conditions.

A new clause 7 should be added to Article 21 requiring employers to develop reasonable-accommodation and return-to-work arrangements for workers who have taken extended leave because of mental health problems where professionally recommended. Measures may include phased return, schedule adjustment, temporary workload reduction, task changes, assignment of a support person, and time off for medical appointments. The cost and scope of accommodation should be proportionate to enterprise size and the nature of the work and should not impose an undue burden.

F. Strengthen dialogue, inspection, and national data

Clause 2 Article 64 of the Labour Code, which lists the subjects the parties may choose for the periodic or on-request workplace dialogue required by clause 2 Article 63, should be amended so that “mental health, psychosocial risks, the use of monitoring technologies, and algorithmic management” become subjects of workplace dialogue when requested by either party. Trade unions and workers’ representative organisations should have access to aggregated information, participate in risk assessments, and monitor remedial plans. Dialogue can prevent a one-size-fits-all right-to-disconnect model while reducing the risk that the rules exist only on paper.

Article 76 of the Law on Occupational Safety and Health should be supplemented to require annual plans to contain objectives and indicators on psychosocial risks where the enterprise falls within a risk category. Inspectorates should be trained to examine work design, working-time data, target-setting systems, harassment complaints, and the impact of technology; inspectors should not require enterprises to provide individual medical records without a lawful basis.

At the national level, the Ministry of Health and the labour-management authority should develop a unified set of indicators, technical guidance, and anonymous reporting mechanisms for work-related stress, sick leave due to psychological problems, violence and harassment, and the impacts of digitalisation. Data should be disaggregated by gender, age, sector, contract type, and mode of work, but must not enable individual identification. Only with reliable data can policy avoid both extremes: underestimating risks and medicalising every workplace conflict.

G. Extend proportionate protection to platform workers

In amending the Labour Code, criteria should be added for presuming an employment relationship where a platform substantially determines prices, work allocation, ratings, access to customers, or account deactivation. Persons meeting the criteria should enjoy rights relating to working time, occupational safety and health, and challenges to algorithmic decisions. For genuinely self-employed persons, the Law on Occupational Safety and Health should still impose minimum duties on platform enterprises: disclose ranking logic, avoid designing incentives that encourage excessive work, allow rest without unreasonable penalties, provide human-review mechanisms, and offer support when violence or harassment occurs.

Conclusion

Digitalisation can improve flexibility, reduce dangerous work, and expand opportunities to participate in the labour market. However, the same technology can create high work intensity, constant connectivity, intrusive monitoring, opaque decision-making, and harassment extending beyond physical spaces. Mental health protection should therefore not be positioned in opposition to innovation; it is a condition for innovation that is sustainable and human-centred.

Vietnamese law already contains constitutional foundations and rules on safety, rest time, dignity, harassment, and personal data, but these elements have not yet formed a coherent psychosocial risk-management cycle. The principal shortcomings are the failure to identify such risks explicitly; the absence of a right to disconnect; the lack of regulation of monitoring and algorithmic management as working-condition issues; and incomplete mechanisms for health examinations, return-to-work support, and the protection of platform workers.

An appropriate reform direction is to integrate mental health into the existing occupational safety and health system while adding enforceable digital rights. The objective is not to prohibit communication, monitoring, or AI, but to require necessity, proportionality, transparency, consultation, control at source, and remedial responsibility. Only when workers can genuinely rest, understand how technology evaluates them, participate in work design, and receive support without stigma can the digital environment become a safe, sustainable, and human-centred work environment.

*****

Footnotes

1. Federico Moja, Digitalisation and Psychosocial Risks, OSHwiki (Eur. Agency for Safety & Health at Work, Feb. 23, 2024), https://oshwiki.osha.europa.eu/en/themes/digitalisation-and-psychosocial-risks (last visited Sept. 3, 2026).

2. World Health Organization, Mental Health at Work (Sept. 15, 2026), https://www.who.int/news-room/fact-sheets/detail/mental-health-at-work.

3. Dao Phu Cuong et al., Some Manifestations of Mental Health Disorders in Workers at Two Production Facilities, 66 Vietnam J. Community Med. (Special Issue 21) 72 (2026), https://tapchiyhcd.vn/index.php/yhcd/article/view/3810.

4. Bui Thi Huyen Trang et al., Insomnia and Mental Health Status Among Staff and Lecturers at Phenikaa University, 66 Vietnam J. Community Med. (Special Issue 10) 1 (2025), https://doi.org/10.52163/yhc.v66iCD10.2604.

5. World Health Organization, Mental Health (Sept. 11, 2026), https://www.who.int/news-room/fact-sheets/detail/mental-health-strengthening-our-response.

6. World Health Organization, supra note 2.

7. Tina Weber & Dragoș Adăscăliței, Eurofound, Right to Disconnect: Implementation and Impact at Company Level 1-5 (Publ’ns Off. Eur. Union 2023).

8. Michele Lerner, Electronically Monitoring Your Employees? It’s Impacting Their Mental Health, Am. Psych. Ass’n (Sept. 7, 2023), https://www.apa.org/topics/healthy-workplaces/employee-electronic-monitoring.

9. Mairi Bowdler et al., Algorithmic Management and Psychosocial Risks at Work: An Emerging Occupational Safety and Health Challenge, 52 Scand. J. Work Env’t & Health 1 (2026), https://doi.org/10.5271/sjweh.4270.

10. Women Win, Understanding Drivers and Prohibitors of Women Garment Workers’ Mental Health & Wellbeing (Feb. 2024), https://www.womenwin.org/win-win-strategies/wp-content/uploads/2024/03/Women-Win_Women-Garment-Workers-Mental-Health-Brief.pdf; see also Women Win, Women Garment Workers’ Mental Health & Wellbeing (Feb. 15, 2024), https://www.womenwin.org/win-win-strategies/news/announcements/women-garment-workers-mental-health-wellbeing/.

11. International Covenant on Economic, Social and Cultural Rights arts. 7, 12, Dec. 16, 1966, 993 U.N.T.S. 3.

12. Occupational Safety and Health Convention art. 4, June 22, 1981, ILO No. 155; Int’l Lab. Org., Ratifications of C155, Normlex, https://normlex.ilo.org/dyn/nrmlx_en/f?p=NORMLEXPUB:11300:0::NO::P11300_INSTRUMENT_ID:312300 (last visited Sept. 3, 2026).

13. World Health Organization, Guidelines on Mental Health at Work 12-61 (2022).

14. World Health Organization & International Labour Organization, Mental Health at Work: Policy Brief 8, 13 (2022).

15. Violence and Harassment Convention art. 3(d), June 21, 2019, ILO No. 190; Int’l Lab. Org., Up-to-Date Conventions and Protocols Not Ratified by Viet Nam, Normlex, https://normlex.ilo.org/dyn/nrmlx_en/f?p=1000:11210:0::NO:11210:P11210_COUNTRY_ID:103004:NO (last visited Sept. 3, 2026).

16. European Parliament Resolution of 21 January 2021 with Recommendations to the Commission on the Right to Disconnect, ¶¶ 13, 16 & annex, recital 10, 2021 O.J. (C 456) 161.

17. Code du Travail [C. trav.] art. L. 2242-17, 7° (Fr.).

18. Tina Weber & Oscar Vargas Llave, Eurofound, Right to Disconnect: Exploring Company Practices 27-54 (Publ’ns Off. Eur. Union 2021).

19. Fair Work Act 2009 (Cth) ss 333M-333N (Austl.); Fair Work Ombudsman, Right to Disconnect (Aug. 26, 2024), https://www.fairwork.gov.au/employment-conditions/right-to-disconnect.

20. Fair Work Commission, What Is the Right to Disconnect (Aug. 26, 2025), https://www.fwc.gov.au/workplace-disputes/other-workplace-disputes/right-disconnect-disputes/what-right-disconnect (last visited Sept. 3, 2026).

21. Hien Phap [Constitution] arts. 21, 35, 38 (2013) (Viet.).

22. Bo Luat Lao Dong [Labor Code], No. 45/2019/QH14, arts. 5(1)(b), 6(2), 63 (2019) (Viet.).

23. Id. arts. 105, 107, 109, 111.

24. Law on Occupational Safety and Health, No. 84/2015/QH13, arts. 7, 18, 21, 76 (2015) (Viet.).

25. Circular No. 19/2016/TT-BYT on Occupational Hygiene Management and Workers’ Health (June 30, 2016) & forms (Viet.).

26. Decree No. 145/2020/ND-CP, art. 84(3) (Dec. 14, 2020) (Viet.).

27. Law on Personal Data Protection, No. 91/2025/QH15, arts. 3, 4, 21, 25, 30, 38(1) (2025) (Viet.).

28. Id. arts. 25(3), 26, 31.

29. Law on Occupational Safety and Health, No. 84/2015/QH13, art. 3(5) (2015) (Viet.).

30. International Labour Organization, ILO and Viet Nam Partners Address Psychosocial Risks at Work in National Occupational Safety and Health Month (May 12, 2026), https://www.ilo.org/resource/news/ilo-and-viet-nam-partners-address-psychosocial-risks-work-national.

31. Labor Code, No. 45/2019/QH14, arts. 105, 107, 118(2)(a) (2019) (Viet.).

32. Tahmina Karimova, AI Systems at Work: A Changing Psychosocial Work Environment 7-18 (Int’l Lab. Org., Working Paper No. 170, 2026), https://doi.org/10.54394/00034098.

33. Law on Occupational Safety and Health, No. 84/2015/QH13, art. 21 (2015) (Viet.).

34. Id. arts. 2(3), 4(4), 6(3)-(4).