(An article by Utkarsh N. Patil, Final year Law student from Government Law College, Mumbai)
ABSTRACT
The institution of standing orders occupies an important position within Indian labour law because it seeks to bring certainty, transparency and uniformity to conditions of employment. The traditional standing-orders framework was developed in an industrial environment characterised by physical workplaces, fixed working hours, direct human supervision and relatively stable categories of employment. Contemporary workplaces, however, increasingly operate through artificial intelligence, algorithmic management, digital surveillance, remote working arrangements and virtual collaboration platforms. These developments have altered not merely the location in which work is performed but also the manner in which managerial authority is exercised.
The transition from the Industrial Employment (Standing Orders) Act 1946 to Chapter IV of the Industrial Relations Code 2020 represents an important restructuring of this legal framework. The Industrial Relations Code now governs standing orders for industrial establishments employing 300 or more workers and provides for Central Government model standing orders, employer-specific standing orders and their certification.¹ The Model Standing Orders notified in 2026 are particularly significant because the service-sector framework expressly recognises work from home, remote locations and virtual workplaces.² Nevertheless, recognition of remote work does not comprehensively address the legal consequences of algorithmic management. Questions concerning automated performance evaluation, workplace surveillance, algorithmic scheduling, AI-assisted disciplinary proceedings, employee privacy, data protection, algorithmic discrimination and the right to meaningful human review remain insufficiently addressed.
This research paper examines the capacity of standing orders to regulate the contemporary AI-enabled workplace. It analyses the historical purpose of standing orders, the statutory framework under the Industrial Relations Code 2020, the implications of the Model Standing Orders 2026, and the relationship between workplace AI and principles of natural justice, privacy, equality and procedural fairness. It further considers the Digital Personal Data Protection Act 2023 and international developments, particularly the European Union Artificial Intelligence Act. The paper argues that standing orders should evolve from instruments concerned primarily with static conditions of service into mechanisms of workplace technology governance. A contemporary standing-orders framework should establish transparency regarding the use of AI, human oversight of consequential employment decisions, safeguards against excessive surveillance, mechanisms for challenging algorithmic decisions, worker consultation and periodic assessment of high-risk workplace technologies.
Keywords: Standing Orders, Industrial Relations Code 2020, Artificial Intelligence, Algorithmic Management, Digital Workplaces, Employee Surveillance, Remote Work, Data Protection, Natural Justice, Labour Law.
1. INTRODUCTION
The employment relationship is fundamentally characterised by an asymmetry of bargaining power. The employer ordinarily possesses greater economic resources, organisational authority and control over the conditions under which work is performed, whereas an individual worker may have limited capacity to negotiate each aspect of the employment relationship. Labour legislation has consequently developed mechanisms intended to limit arbitrary managerial discretion and provide workers with minimum standards of protection. Within this broader framework, standing orders have historically served as an important mechanism for translating general principles of labour protection into specific rules governing conditions of employment.
The Industrial Employment (Standing Orders) Act 1946 was enacted with the objective of requiring employers to define conditions of employment with sufficient precision and to make those conditions known to workmen. The legislative idea was that uncertainty concerning employment conditions could itself facilitate arbitrary managerial action. Standing orders therefore sought to establish an identifiable set of rules governing matters such as classification of workers, attendance, working hours, shifts, leave, suspension, termination and misconduct.
The Supreme Court's jurisprudence demonstrates that standing orders are not merely internal administrative documents. In Agra Electric Supply Co Ltd v Alladin, the Court recognised the importance of certified standing orders in establishing uniform conditions of service.³ In Western India Match Co Ltd v Workmen, the Court held that an employer could not simultaneously maintain statutory standing orders and apply inconsistent contractual conditions to individual employees.⁴ The significance of these decisions extends beyond the industrial workplace in which they arose. Their underlying principle is that conditions of employment should not depend entirely upon unilateral managerial discretion.
The technological transformation of work makes this principle particularly important. A contemporary employee may no longer be supervised exclusively by a human manager. Software can determine which tasks an employee receives, measure the time taken to complete those tasks, monitor computer activity, analyse communications, allocate shifts, rank performance and identify conduct that an employer's system considers suspicious. Artificial intelligence can additionally be used to make predictions concerning employee performance, recruitment suitability, attrition or productivity. Consequently, managerial authority may increasingly be exercised through systems that workers cannot fully observe or understand.
The International Labour Organization has described algorithmic management as the use of algorithmic systems to organise, assign, monitor, supervise and evaluate work.⁵ Such systems are significant because they can change the relationship between the worker and the employer without necessarily changing the formal terms of the employment contract. A worker may technically remain employed under the same contract while experiencing substantial changes in supervision, workload, monitoring and performance evaluation because an algorithm has been introduced into the workplace.
The physical transformation of the workplace is equally important. Remote and hybrid employment arrangements have weakened the traditional distinction between the workplace and the worker's home. Digital communication systems permit employees to work from different geographical locations while remaining continuously connected to organisational systems. The resulting workplace is partly physical and partly digital.
Indian labour law has begun responding to this transformation. Chapter IV of the Industrial Relations Code 2020 deals specifically with standing orders. Section 28 applies the chapter to industrial establishments employing 300 or more workers, while section 29 requires the Central Government to make model standing orders concerning conditions of service and matters incidental thereto.⁶ Section 30 establishes the procedure under which employers prepare draft standing orders based upon the model standing orders, subject to certification.⁷
The Model Standing Orders 2026 constitute a further development because the service-sector framework expressly recognises work from home, work from a remote location and work in a virtual workplace.⁸ This is an important departure from a conception of employment centred exclusively upon a physical establishment. However, the recognition of remote work raises a second-order legal question: if the workplace has become digital, should the standing orders also regulate the digital mechanisms through which work is supervised?
This paper proceeds on the premise that they should. If an employer can exercise substantial managerial power through an algorithm, then the principles underlying standing orders require that such power be subject to transparent and legally reviewable rules.
2. HISTORICAL PURPOSE AND LEGAL CHARACTER OF STANDING ORDERS
The historical justification for standing orders lies in the need to make conditions of employment certain. Before statutory standing orders, important matters concerning employment could be determined through managerial practice, individual contracts or informal workplace rules. Such an arrangement could produce considerable uncertainty for workers, particularly where the employer retained broad discretion concerning discipline, attendance, termination or other conditions of service.
The Industrial Employment (Standing Orders) Act 1946 attempted to address this problem by requiring employers to formally define conditions of employment and communicate them to workers. The legislation consequently represented a movement away from purely contractual employment towards a regulated employment relationship in which certain terms were required to be objectively established.
The importance of standing orders is evident from the Supreme Court's interpretation of their legal character. In Agra Electric Supply Co Ltd v Alladin, the Supreme Court considered the binding effect of certified standing orders and recognised their role in establishing uniformity in employment conditions.⁹ In Western India Match Co Ltd v Workmen, the Court held that the employer could not maintain conditions inconsistent with the certified standing orders merely through individual contractual arrangements.¹⁰ The significance of this principle lies in its recognition that an employer cannot avoid statutory regulation simply by shifting the relevant condition into another contractual or administrative instrument.
This principle has considerable relevance to digital workplaces. Suppose an employer's certified standing orders regulate attendance and disciplinary action but an internal AI system automatically determines that an employee has breached the attendance requirements. If the algorithm effectively establishes a new attendance regime that differs from the standing orders, the employer would be using technology to create a parallel set of employment rules. The legal issue would not disappear merely because the rule is implemented through software.
Standing orders should therefore be understood as instruments controlling the exercise of managerial power. Their importance does not depend upon whether management is exercised by a supervisor, a human-resources department or an AI-enabled system.
3. THE INDUSTRIAL RELATIONS CODE 2020 AND THE NEW STANDING-ORDERS FRAMEWORK
The Industrial Relations Code 2020 consolidates the legal framework relating to trade unions, standing orders and industrial disputes. Chapter IV specifically addresses standing orders and represents the current statutory foundation of this area of Indian labour law. Section 28 provides that the chapter applies to industrial establishments employing 300 or more workers or having employed that number on any day during the preceding twelve months.¹¹
The threshold is significant because standing orders impose a formal regulatory burden on employers. The legislature has consequently adopted a numerical threshold intended to distinguish establishments for which formal standing orders are considered necessary from smaller establishments. However, the threshold becomes more complicated in the digital economy. A technology company may exercise extensive algorithmic control over workers while employing fewer than 300 direct employees, particularly where outsourcing, contractual employment and platform arrangements are used.
Section 29 requires the Central Government to make model standing orders relating to conditions of service and matters incidental thereto.¹² The model standing orders serve as a statutory template against which employer-specific standing orders may be developed. Section 30 requires the employer to prepare draft standing orders based upon the model standing orders while permitting provisions appropriate to the particular establishment, provided that they remain consistent with the Code.¹³
The certification process is important in the context of AI because it potentially provides a mechanism for evaluating the fairness and reasonableness of technologically mediated employment rules. If an employer proposes a standing order allowing continuous employee surveillance, automated performance scoring and disciplinary action based upon algorithmic outputs, the question should not merely be whether such technology is technically possible. The relevant legal inquiry should also include whether the proposed employment condition is fair and reasonable.
The Code also provides that standing orders are to be made available to workers and establishes a statutory mechanism concerning their operation, modification and interpretation.¹⁴ This reinforces the principle that employment rules must remain accessible to those who are governed by them.
The technological challenge is therefore not necessarily the absence of a legal mechanism. Rather, the difficulty lies in applying a framework developed for relatively conventional employment conditions to technologies capable of producing complex and opaque forms of managerial control.
4. THE MODEL STANDING ORDERS 2026 AND THE DIGITAL WORKPLACE
The Model Standing Orders 2026 represent a particularly important development for the purposes of this research. The service-sector standing orders expressly recognise work from home, work from a remote location and work in a virtual workplace.¹⁵ This recognition reflects the changing geography of employment and acknowledges that the performance of work need not occur within the employer's physical premises.
The legal significance of this provision extends beyond convenience. Traditional labour law often assumed that the workplace was a physically identifiable establishment where the employer exercised supervision. Remote work disrupts this assumption. An employee may work from home, another city or another country while remaining connected to the employer through digital systems.
The recognition of virtual workplaces therefore provides a basis for extending the concept of workplace regulation to digital environments. However, the provision primarily recognises the location from which work may be performed. It does not by itself create a comprehensive framework governing the technologies used to supervise remote workers.
This distinction is important. An employee working from home may be monitored through login data, screen-monitoring software, webcam systems, keystroke monitoring, productivity dashboards or AI-based behavioural analysis. The fact that the employee has been permitted to work remotely does not answer whether these forms of monitoring are lawful, necessary or proportionate.
Consequently, the Model Standing Orders 2026 should be understood as an important foundation for digital workplace regulation rather than a complete solution to the challenges created by AI.
5. ARTIFICIAL INTELLIGENCE AND ALGORITHMIC MANAGEMENT
Artificial intelligence is increasingly incorporated into employment systems at different stages of the employment relationship. Employers may use AI during recruitment, onboarding, scheduling, performance evaluation, workforce planning and disciplinary processes. Algorithmic systems may also be used without technically qualifying as artificial intelligence, particularly where rules-based software performs managerial functions.
The labour-law significance of these systems lies in their capacity to alter the exercise of managerial authority. A human manager traditionally makes a decision after observing a worker and exercising personal judgment. An algorithm may instead produce a numerical score or recommendation based upon a large volume of data. Although the resulting decision may appear objective, the algorithm reflects choices made by its developers and the employer concerning what data should be collected, what variables should be considered and what outcomes should be treated as desirable.
The use of algorithmic management therefore creates a potential distinction between formal managerial responsibility and functional managerial power. The employer may formally remain responsible for an employment decision, while the substantive decision-making process has been delegated to software.
This is particularly significant for standing orders. If standing orders exist partly to prevent arbitrary exercise of managerial discretion, then the delegation of managerial discretion to algorithms should not remove that activity from their regulatory scope.
6. AI-BASED RECRUITMENT AND EMPLOYMENT DECISIONS
Artificial intelligence may be used during recruitment to screen applications, rank candidates, analyse assessments and identify characteristics associated with successful employees. These systems can reduce administrative burdens where employers receive large numbers of applications. At the same time, they can reproduce patterns contained within historical employment data.
If a recruitment algorithm is trained on historical hiring decisions, it may reproduce the preferences or biases contained within those decisions. The resulting discrimination may be difficult to identify because the algorithm may not explicitly use a protected characteristic but may rely upon variables that operate as proxies for it.
The European Union has recognised the particular significance of employment-related AI. Regulation (EU) 2024/1689 classifies certain AI systems used in employment and worker management as high-risk, including systems used for recruitment and selection, decisions affecting employment relationships, promotion and termination, task allocation and monitoring or evaluation of workers.¹⁶
This comparative approach is relevant to India because recruitment and employment decisions directly affect livelihood and economic opportunity. A standing-orders framework responsive to AI should therefore require employers to disclose the material use of automated systems in employment decisions and ensure that consequential decisions are subject to meaningful human review.
7. ALGORITHMIC PERFORMANCE MANAGEMENT
Performance management represents perhaps the most significant area in which AI can alter traditional employment relationships. A conventional performance assessment may be based upon supervisor observations, output, attendance and periodic reviews. Digital systems can expand this process by collecting extensive behavioural data.
An employee's performance may be assessed through the number of tasks completed, response times, login duration, customer ratings, communication patterns, error rates and other metrics. AI systems may combine these data points into a single performance score.
The apparent mathematical precision of such systems can create the impression that algorithmic assessments are inherently objective. However, an algorithm does not eliminate subjectivity; it can relocate subjectivity into the design of the system. The choice of variables, weighting of indicators and definition of successful performance remain managerial decisions.
The ILO has highlighted concerns regarding algorithmic management and the quality of the objectives, data and programming underlying AI-enabled human-resource systems.¹⁷ This suggests that standing orders should not simply recognise an employer's right to monitor performance. They should establish procedural safeguards concerning how performance information is generated and used.
A worker should have an opportunity to challenge an inaccurate performance record. For example, an employee might appear inactive because of a network failure, might receive fewer tasks because of an algorithmic scheduling error or might be penalised because customer ratings are systematically affected by factors outside the employee's control. Treating the algorithmic output as conclusive would convert a potentially fallible technological indicator into a disciplinary fact.
8. WORKPLACE SURVEILLANCE AND EMPLOYEE PRIVACY
Digital workplaces significantly increase employers' capacity to monitor workers. Modern surveillance can include biometric attendance, location tracking, screen monitoring, email analysis, browser monitoring, CCTV, facial recognition and AI-based behavioural analysis.
Employee surveillance raises a fundamental question concerning the balance between the employer's legitimate interests and the worker's privacy and dignity.
The Supreme Court's decision in Justice KS Puttaswamy (Retd) v Union of India recognised privacy as a constitutionally protected right under article 21.¹⁸ The Court's later jurisprudence also developed the requirements of legality, legitimate state purpose and proportionality when assessing restrictions upon privacy.¹⁹ Although constitutional rights operate primarily against the State, the principles articulated by the Supreme Court have wider significance for employment law, particularly where legislation and contractual principles regulate private employment relationships.
The emergence of digital surveillance demonstrates why workplace privacy cannot be reduced to the question of whether an employee has a "private" physical space. A worker performing employment duties from home may simultaneously be inside a private residence and connected to the employer's digital infrastructure. Continuous monitoring can therefore create a form of workplace intrusion extending beyond the employer's physical premises.
Standing orders should accordingly identify the purposes for which employee monitoring may be undertaken, the categories of data that may be collected, the persons authorised to access such information, the retention period and the circumstances in which monitoring may be used for disciplinary purposes.
9. DATA PROTECTION AND STANDING ORDERS
The Digital Personal Data Protection Act 2023 establishes India's statutory framework for processing digital personal data.²⁰ The relevance of the legislation to employment is considerable because employers routinely process large quantities of employee information.
The digital workplace may generate information concerning attendance, location, communications, performance, access to computer systems and other aspects of employment. AI systems can combine such information to generate predictions or classifications concerning workers.
Data-protection legislation and standing orders should therefore be understood as complementary regulatory mechanisms. Data-protection law establishes general rules concerning processing of digital personal data, while standing orders can regulate the employment-specific circumstances in which workplace technologies are deployed.
For example, standing orders could require an employer to disclose that an AI system is used to evaluate performance, identify the categories of information used by that system and establish a procedure through which an employee may challenge an inaccurate outcome.
This would create a connection between data governance and labour governance. The central objective should be to ensure that employees do not lose substantive workplace rights merely because the information used to make a decision has been processed digitally.
10. AI, DISCIPLINARY PROCEEDINGS AND NATURAL JUSTICE
The use of AI in disciplinary proceedings presents one of the most difficult questions for labour law. Traditional disciplinary proceedings generally involve an identifiable allegation, evidence supporting the allegation, an opportunity for the worker to respond and a decision by an authorised person.
Algorithmic management can disrupt this sequence. A system may automatically identify an employee as having violated a rule and generate an alert that subsequently becomes the basis for disciplinary proceedings.
The legal difficulty is that an algorithmic output is not necessarily equivalent to established misconduct. It may constitute evidence requiring verification.
The Supreme Court's decision in Workmen of Firestone Tyre & Rubber Co of India (P) Ltd v Management demonstrates the importance of procedural fairness in domestic disciplinary proceedings.²¹ Similarly, in DK Yadav v JMA Industries Ltd, the Supreme Court emphasised the relationship between termination of employment, livelihood and fair procedure.²²
These principles should apply irrespective of whether the original allegation was generated by a human supervisor or an algorithm.
Consider a situation in which an AI system identifies an employee as having falsified attendance because the system records no activity during a particular period. If the employee was actually performing work offline, attending a physical meeting or experiencing a network failure, the algorithmic conclusion would be inaccurate. A standing order that allows immediate disciplinary action based solely on the algorithm would effectively eliminate the worker's opportunity to contest the factual basis of the allegation.
The appropriate approach is therefore to require meaningful human review before consequential disciplinary action is taken.
11. THE PRINCIPLE OF HUMAN OVERSIGHT
Human oversight should not be reduced to the formal presence of a manager who automatically approves an algorithmic recommendation. Meaningful oversight requires the decision-maker to possess sufficient authority and information to question the technological output.
A modern standing order should therefore establish that AI-generated recommendations concerning termination, suspension, disciplinary penalties, promotion, demotion or significant compensation decisions cannot be treated as conclusive.
The responsible decision-maker should examine the relevant circumstances independently and provide reasons for the final decision.
This principle is particularly important because the use of AI can create what may be described as automation bias: the tendency to treat computer-generated conclusions as inherently more objective or accurate than human judgment. In employment law, such bias can have serious consequences because employment decisions directly affect livelihood.
Human oversight therefore serves two purposes. First, it provides an opportunity to identify technical errors. Secondly, it preserves accountability by ensuring that an identifiable human decision-maker remains responsible for consequential employment action.
12. THE RIGHT TO AN EXPLANATION
An additional problem is the explainability of AI systems. A traditional managerial decision can generally be expressed through reasons such as poor attendance, inadequate performance or violation of a workplace rule. An algorithm may instead produce a numerical risk score without explaining precisely how that score was generated.
A worker challenging such a decision requires sufficient information to determine whether the underlying data were accurate and whether the decision was based upon appropriate criteria.
A standing order need not necessarily require employers to disclose proprietary source code. It should, however, require disclosure of the purpose of the AI system, the principal categories of information used, the nature of the decision influenced by the system and whether human review occurred.
Such a limited right to explanation would promote procedural fairness without necessarily requiring employers to disclose commercially sensitive technical information.
13. AI, EQUALITY AND ALGORITHMIC DISCRIMINATION
Equality concerns are particularly important because AI systems may reproduce historical patterns of discrimination. An algorithm trained on previous employment decisions may learn from patterns that reflect historical inequalities.
The problem can occur even where the algorithm does not expressly use a protected characteristic. Variables such as educational institution, geographical location, employment history or patterns of communication may operate as proxies for characteristics that produce discriminatory outcomes.
Employment-related AI is therefore appropriately treated as a high-impact category in comparative regulation. The EU AI Act's classification of employment-related systems as high-risk reflects the potentially significant consequences of automated employment decisions.²³
Indian standing orders should respond to this problem through periodic review of high-impact AI systems. Employers deploying AI for recruitment, promotion, performance evaluation or termination should be required to examine whether the system produces systematically discriminatory outcomes.
14. REMOTE WORK AND THE RIGHT TO DISCONNECT
The formal recognition of remote work by the Model Standing Orders 2026 is significant, but remote work creates additional questions concerning working time.
Digital connectivity allows employees to receive emails, messages and work assignments outside conventional working hours. A worker may therefore remain physically outside the workplace while continuing to experience managerial expectations through digital systems.
This raises the possibility of an "always-connected" workplace in which formal working hours and actual availability diverge.
Standing orders should consequently specify working hours, permitted communication outside working hours, emergency exceptions and expectations regarding response times. They should also establish that employees should not ordinarily face adverse consequences for refusing to perform work outside prescribed hours unless the employment arrangement lawfully requires such availability.
The right to disconnect is therefore closely connected with the traditional labour-law regulation of working time. Digital technology should not make working-time protections practically meaningless.
15. DIGITAL HARASSMENT AND THE EXPANDED CONCEPT OF WORKPLACE
The Supreme Court's decision in Vishaka v State of Rajasthan demonstrates that workplace rules must respond to the actual forms through which workplace harm occurs.²⁴ The Court recognised the need for preventive mechanisms against sexual harassment and contemplated incorporation of relevant safeguards into standing orders.
The contemporary workplace requires the concept of workplace harassment to include digital environments. Employees may experience harassment through email, messaging platforms, video-conferencing systems and other workplace communication tools. Artificial intelligence additionally creates new risks, including the creation or dissemination of manipulated images and other synthetic content.
The legal principle underlying Vishaka is therefore adaptable to the digital workplace. The relevant question should not be whether misconduct occurred inside a physical office but whether it occurred within a work-related environment or through systems connected to employment.
Modern standing orders should expressly clarify that authorised digital workplace platforms constitute part of the workplace for purposes of disciplinary and anti-harassment rules.
16. GENERATIVE AI AND EMPLOYEE CONDUCT
The development of generative AI has created a new category of workplace conduct requiring regulation. Employees may use AI tools to prepare documents, write software, analyse information, translate material or automate routine tasks. At the same time, inappropriate use of such systems can create confidentiality, cybersecurity, accuracy and intellectual-property risks.
A blanket prohibition on generative AI may be impractical. A more appropriate approach would distinguish between legitimate and prohibited uses.
Standing orders could establish that ordinary low-risk use of approved AI tools is permitted, while use involving confidential information, personal data or sensitive corporate information requires authorisation. The unauthorised transfer of confidential information to an external AI system could constitute misconduct where the worker knowingly violates established rules.
This approach would also require employers to provide clear instructions. An employee should not ordinarily be disciplined for violating an AI policy that was never communicated or sufficiently explained.
17. AI-GENERATED EVIDENCE AND WORKPLACE INVESTIGATIONS
Digital workplaces generate substantial quantities of evidence. Access logs, emails, CCTV recordings, metadata and AI-generated alerts may all become relevant during disciplinary investigations.
The existence of such evidence does not eliminate the need to assess its reliability.
An AI-generated alert should therefore be treated as an evidentiary lead rather than automatically as proof of misconduct. The employer should preserve the underlying records necessary to verify the allegation and should provide the worker with sufficient information to respond.
This approach is consistent with the broader principle of natural justice. A disciplinary system becomes procedurally defective if the worker is required to answer an allegation without knowing the factual basis upon which it rests.
The use of AI therefore increases, rather than decreases, the importance of clear disciplinary procedures within standing orders.
18. ALGORITHMIC MANAGEMENT AND WORK INTENSIFICATION
AI can increase work intensity even when it does not replace human workers. An algorithm designed to maximise productivity may continuously adjust task allocation, expected completion times and performance targets.
A worker may consequently experience a significant increase in the amount or speed of work expected from them without any formal change to their job title or contractual position.
This creates a difficult regulatory problem because traditional labour law often focuses upon measurable conditions such as wages, hours and physical safety. Algorithmic work intensification can occur through apparently neutral performance metrics.
The ILO has identified concerns regarding the impact of algorithmic management on working conditions and job quality.²⁵ Standing orders should therefore address the consequences of digital performance management rather than treating productivity software as a purely managerial or technological matter.
19. TRADE UNIONS AND WORKER PARTICIPATION IN AI GOVERNANCE
The introduction of AI can affect not only individual employees but also collective employment conditions. Algorithmic scheduling may change shift patterns; automated performance management may change evaluation standards; automation may reduce demand for particular skills; and digital monitoring may alter the balance of power between workers and management. Worker participation is therefore essential.
The Industrial Relations Code's standing-orders framework already incorporates a consultative dimension through the involvement of the negotiating union or negotiating council in the standing-orders process.²⁶ This provides a potential institutional basis for extending worker participation to technological changes.
Worker representatives should have an opportunity to understand the employment consequences of significant AI systems before those systems are deployed. Such consultation should include the purpose of the system, categories of workers affected, monitoring implications, anticipated changes in work organisation and safeguards against discrimination or inaccurate decision-making.
The objective should not be to give workers a veto over technological innovation. Rather, consultation should ensure that technological transformation remains integrated into the collective regulation of employment.
20. COMPARATIVE PERSPECTIVE: THE EUROPEAN UNION AI ACT
The European Union Artificial Intelligence Act provides one of the most developed comparative approaches to workplace AI. Regulation (EU) 2024/1689 classifies specified AI systems used in employment and worker management as high-risk. The relevant category includes systems used for recruitment and selection, decisions affecting employment relationships, promotion and termination, task allocation and monitoring or evaluation of workers.²⁷
The significance of this approach lies in its risk-based methodology. It does not treat all AI systems as equally dangerous. Instead, regulatory obligations increase according to the potential consequences of the technology.
A similar principle could be incorporated into Indian standing orders. Routine administrative AI systems could be subject to relatively limited disclosure requirements, while systems affecting recruitment, termination, discipline, compensation or employee surveillance could be subject to enhanced safeguards.
The comparative experience also demonstrates that AI regulation need not prevent technological innovation. Rather, it can establish governance requirements proportionate to the consequences of the technology.
21. MAJOR GAPS IN THE INDIAN STANDING-ORDERS FRAMEWORK
The current Indian framework provides a statutory foundation but does not yet constitute a comprehensive system of AI governance at the workplace.
The first significant gap is the absence of a comprehensive statutory definition of algorithmic management. Without such a definition, it may be unclear which technological systems fall within the regulatory framework.
The second gap concerns transparency. Workers may not necessarily know when an AI system is being used to evaluate or monitor them.
The third concerns automated decision-making. The present standing-orders framework does not establish a comprehensive right to human review where an algorithm materially influences a consequential employment decision.
The fourth concerns surveillance. Remote work recognition does not itself determine when continuous digital monitoring is justified.
The fifth concerns algorithmic discrimination. There is no comprehensive standing-orders requirement for employers to test high-impact employment AI systems for discriminatory outcomes.
The sixth concerns the role of worker representatives. Existing consultation mechanisms provide a foundation but do not specifically establish technological consultation before the introduction of high-impact AI.
Finally, the numerical threshold applicable to Chapter IV raises concerns regarding smaller technology-intensive establishments. A workplace with fewer than 300 workers may nevertheless exercise extensive algorithmic control over its employees.
22. PROPOSED FRAMEWORK FOR AI-RESPONSIVE STANDING ORDERS
The future model standing orders should expressly regulate artificial intelligence and algorithmic management. Such provisions should not be drafted merely as an IT policy. They should be integrated into the legal framework governing conditions of employment.
The first requirement should be transparency. Employers should inform workers where AI is materially involved in recruitment, performance evaluation, task allocation, scheduling, monitoring or disciplinary decision-making. Employees should know whether the decision affecting them has been influenced by an automated system.
The second requirement should be meaningful human oversight. A decision involving dismissal, suspension, disciplinary punishment, promotion, demotion or significant compensation should not be based solely upon an automated output. A designated human decision-maker should independently review the relevant circumstances and remain accountable for the final decision.
The third requirement should be an employee's right to challenge consequential algorithmic decisions. Where an AI-generated assessment materially affects employment, the worker should be able to seek review through the employer's grievance or disciplinary mechanism.
The fourth requirement should be proportionality in workplace surveillance. Employers should identify the legitimate purpose for which monitoring is undertaken and should use the least intrusive mechanism reasonably capable of achieving that purpose.
The fifth requirement should be algorithmic impact assessment. Before deploying high-risk AI systems, employers should examine potential effects upon privacy, equality, employment security, work intensity and procedural fairness.
The sixth requirement should be worker consultation. Where a proposed technology is likely to materially alter working conditions, employee representatives should be consulted before implementation.
The seventh requirement should concern data governance. Standing orders should specify categories of employee data collected for workplace AI, purposes of collection, access rights, retention periods and security measures, subject to the applicable data-protection framework.
23. THE FUTURE ROLE OF THE CERTIFYING OFFICER
The technological transformation of employment also raises questions concerning the institutional capacity of labour authorities. The certifying officer traditionally evaluates whether standing orders comply with statutory requirements and whether their provisions are fair and reasonable.
In the AI era, assessing fairness may require understanding the operation and consequences of technological systems.
The solution need not be to transform labour authorities into general technology regulators. Instead, certifying officers could be provided with appropriate guidance and access to technical expertise where standing orders involve high-impact surveillance or algorithmic management.
For example, where an employer seeks to introduce standing orders authorising AI-based employee monitoring, the certifying authority should be capable of examining the purpose of the monitoring, the categories of information collected and the procedural safeguards available to employees.
This would allow the existing labour-law framework to accommodate new technologies without creating an entirely separate regulatory institution.
24. RESKILLING, AUTOMATION AND EMPLOYMENT SECURITY
Artificial intelligence creates another challenge for labour law because technological transformation can alter the demand for particular forms of labour. Some tasks may become automated, while other occupations may be transformed rather than eliminated.
The ILO's research on generative AI has emphasised the distinction between job transformation and complete job displacement.²⁸ This distinction is important because labour law should not treat every technological change as an individual failure of employee performance.
Standing orders should distinguish between misconduct, poor performance, technological redundancy and restructuring. An employee should not be subjected to disciplinary action merely because technology has reduced the economic need for a particular task.
Where feasible, employers introducing substantial automation should consider reskilling, training and redeployment opportunities. This would allow standing orders to contribute to a more orderly technological transition.
25. BALANCING EMPLOYER INTERESTS AND WORKER RIGHTS
The regulation of AI in the workplace should not proceed from the assumption that technological systems are inherently harmful. Employers have legitimate interests in improving productivity, preventing fraud, protecting confidential information, maintaining cybersecurity and ensuring workplace safety.
Workers, on the other hand, have legitimate interests in privacy, dignity, equality, fair evaluation, employment security and procedural fairness.
The appropriate regulatory approach is therefore not a prohibition on workplace AI but a proportionality-based framework.
The central question should be whether the technological intervention is reasonably connected to a legitimate employment purpose and whether its impact upon workers is appropriately controlled.
Standing orders are particularly suitable for this balancing exercise because they operate at the level at which general labour-law principles become concrete workplace rules.
26. FINDINGS OF THE STUDY
The analysis demonstrates that standing orders remain highly relevant in the age of artificial intelligence. Their importance arises precisely because AI can increase the scale and opacity of managerial decision-making. Where a traditional supervisor could make an arbitrary decision affecting one employee, an algorithmic system may apply the same decision-making logic to thousands of employees simultaneously.
The Industrial Relations Code 2020 provides the institutional structure necessary for continuing the standing-orders system. The Model Standing Orders 2026 further demonstrate adaptation to changing forms of work by recognising remote and virtual workplaces.²⁹ However, the regulatory framework has not yet developed equivalent detail concerning AI-enabled decision-making.
The principal challenge is therefore not whether standing orders remain relevant, but whether their traditional concepts of fairness, reasonableness, transparency and procedural protection can be translated into a technologically mediated workplace.
The analysis suggests that they can. Indeed, the principles underlying standing orders are sufficiently broad to accommodate technological change, provided that the rules themselves are updated.
27. RECOMMENDATIONS
The Indian standing-orders framework should expressly recognise artificial intelligence and algorithmic management as matters capable of affecting conditions of employment. Employers using AI for consequential employment decisions should be required to disclose its material use to affected workers.
High-impact employment decisions should be subject to meaningful human review. An AI system should not become a substitute for managerial accountability or procedural fairness.
Standing orders should establish safeguards for workplace surveillance, including purpose limitation, proportionality, data minimisation and appropriate retention periods. Workers should be informed of significant monitoring systems and should have a mechanism to challenge inaccurate information generated by those systems.
Employers should conduct impact assessments before deploying AI systems capable of materially affecting employment. These assessments should consider privacy, discrimination, accuracy, work intensity, cybersecurity and procedural fairness.
Worker representatives should be consulted before significant AI systems are introduced where they are likely to alter working conditions. The objective should be technological adaptation through social dialogue rather than technological change occurring entirely outside the industrial-relations framework.
The labour administration should also develop specialised guidance concerning AI-enabled standing orders. Certifying officers should have access to appropriate technological expertise where necessary to evaluate complex provisions concerning algorithmic management.
Finally, standing orders should incorporate a clear distinction between technological redundancy and employee misconduct. The introduction of automation should not itself become a basis for disciplinary action against workers whose roles are affected by technological change.
28. CONCLUSION
Standing orders were historically developed to solve a fundamental problem in labour relations: the uncertainty created when significant employment conditions are determined primarily through unilateral managerial discretion. The technological transformation of work has not eliminated that problem. It has changed its form.
The traditional supervisor has increasingly been supplemented by algorithms. The physical factory has increasingly been supplemented by the virtual workplace. The attendance register has increasingly been supplemented by digital activity records. Human performance assessments have increasingly been supplemented by algorithmic scores. These developments make the underlying purpose of standing orders more important rather than less important.
The Industrial Relations Code 2020 provides the current statutory foundation for standing orders in India, while the Model Standing Orders 2026 demonstrate an important recognition of remote and virtual forms of work.³⁰ Yet the regulation of digital workplaces cannot stop at recognising where work is performed. It must also address how work is monitored, evaluated and controlled.
Artificial intelligence presents a particular challenge because it can make managerial authority less visible. A worker may know that a decision has been made but may not know how the relevant algorithm reached its conclusion. The legal response should therefore be to preserve the fundamental principles that have historically justified standing orders: certainty, transparency, fairness and accountability.
A future standing-orders framework should accordingly recognise that technological neutrality does not mean regulatory neutrality. An algorithm should not acquire greater legal authority merely because it is technological. If a human manager could not lawfully make a particular employment decision without following prescribed procedures, the employer should not be able to avoid those procedures simply by transferring the decision-making process to software.
The appropriate future of standing orders is therefore not their disappearance but their transformation. Standing orders should evolve into instruments of workplace technology governance, integrating labour-law principles with rules concerning artificial intelligence, employee data, digital surveillance, remote work and algorithmic decision-making.
The fundamental principle should remain unchanged: technological transformation may change the way managerial power is exercised, but it should not remove that power from the discipline of labour law.
FOOTNOTES
1. Industrial Relations Code 2020, ss 28–30.
2. Ministry of Labour and Employment, Model Standing Orders for Service Sector (2026) para 9.
3. Agra Electric Supply Co Ltd v Alladin (1970) 2 SCC 598.
4. Western India Match Co Ltd v Workmen (1974) 3 SCC 330.
5. International Labour Organization, Algorithmic Management in the Workplace (ILO 2024).
6. Industrial Relations Code 2020, ss 28–29.
7. ibid s 30.
8. Ministry of Labour and Employment, Model Standing Orders for Service Sector (2026) para 9.
9. Agra Electric Supply Co Ltd v Alladin (n 3).
10. Western India Match Co Ltd v Workmen (n 4).
11. Industrial Relations Code 2020, s 28(1).
12. ibid s 29(1).
13. ibid s 30(1).
14. ibid ss 33–37. The Code expressly provides for the operation, availability, registration, modification and interpretation of standing orders.
15. Ministry of Labour and Employment, Model Standing Orders for Service Sector (2026) para 9.
16. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence, annex III.
17. Janine Berg and Hannah Johnston, ‘AI in Human Resource Management: The Limits of Empiricism’ (ILO Working Paper No 154, 2025).
18. Justice KS Puttaswamy (Retd) v Union of India (2017) 10 SCC 1.
19. KS Puttaswamy (Retd) v Union of India (2019) 1 SCC 1.
20. Digital Personal Data Protection Act 2023.
21. Workmen of M/s Firestone Tyre & Rubber Co of India (P) Ltd v Management (1973) 1 SCC 813.
22. DK Yadav v JMA Industries Ltd (1993) 3 SCC 259.
23. Regulation (EU) 2024/1689, annex III.
24. Vishaka v State of Rajasthan (1997) 6 SCC 241.
25. Uma Rani, Annarosa Pesole and Ignacio Gonzalez Vazquez, Algorithmic Management Practices in Regular Workplaces: Case Studies in Logistics and Healthcare (ILO and European Commission Joint Research Centre 2024).
26. Industrial Relations Code 2020, s 30.
27. Regulation (EU) 2024/1689, annex III.
28. International Labour Organization, Generative AI and Jobs: A 2025 Update (ILO 2025).
29. Ministry of Labour and Employment, Model Standing Orders for Service Sector (2026) para 9.
30. Industrial Relations Code 2020, ch IV.
