https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 DOI: https://doi.
org/10.
38035/gijlss.
https://creativecommons.
org/licenses/by/4.
Settlement of Industrial Relations Disputes Based on Work Decisions Generated by Automatic Systems Firkha Amidana1* .
Adi Muliawansyah Malie 2 Universitas Negeri Surabaya.
Jawa Timur.
Indonesia, firkha.
23095@mhs.
Universitas Negeri Surabaya.
Jawa Timur.
Indonesia, adimalie@unesa.
*Corresponding Author: firkha.
23095@mhs.
Abstract: The development of information technology encourages the use of automated systems in employment decision-making, such as performance assessment, supervision, sanctions, transfers, and the sustainability of employment relationships.
While it improves efficiency, it has the potential to lead to industrial relations disputes, especially if done without transparency, correction mechanisms, and adequate human oversight.
This study aims to analyze the legal status of automated system-based work decisions and dispute resolution in Indonesian law, especially in the case of worker mutation to remote locations to encourage resignation and avoid severance pay obligations.
The method used is normative legal research with legislative, conceptual, and case approaches.
The results of the study show that automated systems are not subject to law, so the responsibility remains with the Decision testing refers to the principles of legal protection, legal certainty, procedural fairness, and accountability.
Keywords: Industrial Relations.
Automated Systems.
Work Decisions.
Worker Transfers.
Severance Pay.
INTRODUCTION
The rapid advancement of information technology has fundamentally reconfigured the governance of labor relations, particularly through the emergence of algorithmic management These systems, as defined by the International Labour Organization, utilize datadriven mechanisms to organize, assign, monitor, and evaluate work processes (International Labour Organization, 2.
Notably, algorithmic management is not confined to advanced artificial intelligence.
even relatively simple rule-based systems may qualify as such, insofar as they materially shape managerial decisions and affect workersAo conditions (Annarosa Pesole et al.
This broader understanding shifts the analytical focus from technological sophistication to the extent of its impact on workersAo rights and employment outcomes.
From a managerial perspective, automated decision-making systems offer clear advantages, including efficiency gains, cost reduction, and procedural consistency.
However, within the context of industrial relations, such efficiency-oriented justifications are normatively insufficient.
Employment-related decisions inherently implicate workersAo 437 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 dignity, economic security, and social welfare, thereby requiring alignment with constitutional guarantees of legal protection and fair treatment (Undang-Undang Dasar Negara Republik Indonesia 1.
This concern is further exacerbated by the structurally unequal nature of employment relationships, where employersAo control over capital, organizational design, and increasingly, digital infrastructures and data ecosystems, places workers in a comparatively vulnerable position (Satjipto Rahardjo, n.
A particularly critical manifestation of this issue arises when automated performance evaluation systems are operationalized as the basis for consequential human resource One such practice is the reassignment .
of workers to geographically distant locations.
Although formally justified as organizational restructuring or workforce redistribution, such measures may impose disproportionate burdens on workers, including increased living costs, disruption of family life, and psychological distress (Angling Dandy Wicaksana 2.
In more problematic instances, these reassignment policies function as indirect mechanisms of constructive dismissal, strategically designed to induce voluntary resignation and thereby circumvent employersAo obligations to provide severance pay (Intan Rakhmawati 2.
That such a pattern is not just a theoretical possibility can be seen from several cases that have emerged.
In the study of the Industrial Relations Court Decision Number 155/Pdt.
Sus-PHI/2019/PN.
JKT.
PST (AuIndustrial Relations Court Decision No.
155/Pdt.
SusPHI/2019/PN.
JKT.
PSTAy 2.
the ruling noted that mutations in several cases are often abused by employers to avoid the obligation to pay severance pay, usually by moving workers to distant places or to lower positions so that workers resign.
These findings are important because they show that mutations are not always neutral as managerial policies, but can be a tool of disguised pressure in working relationships.
This means that when a performance system is used to choose who to mutate, it can be part of a mechanism that substantively suppresses workers out of the company (Angling Dandy Wicaksana 2.
In addition, problematic mutation practices are also seen in court rulings.
In the Palu District Court Decision Number 53/Pdt.
Sus-PHI/2018/PN Pal, the court stated that the mutation letter, the first summons, and the second summons against the plaintiff were invalid and null and void, and stated that the termination of employment carried out by the defendant was also invalid and null and void.
This ruling shows that in the practice of industrial relations justice, the mutation policy is not automatically considered valid just because it comes from the company's management.
If the mutation or derivative steps are contrary to labor law and detrimental to workers, then the court can consider the action to be legally From this, it can be seen that the industrial relations dispute forum actually has space to assess the substance of the mutation policy, not just the formal form of the decision.
The presumed objectivity of automated systems also warrants critical examination.
Algorithmic decision-making processes are inherently shaped by human choices regarding data selection, performance indicators, and evaluative criteria.
Consequently, such systems are not value-neutral but reflect embedded institutional preferences and biases.
The International Labour Organization further notes that algorithmic management reduces direct interaction between workers and human decision-makers, thereby increasing the opacity of decision-making processes and the risk of unjust outcomes (Kathleen C.
Kellogg and Valentine 2.
In such contexts, workers are often excluded from meaningful access to the data and reasoning that underpin decisions affecting their employment status.
Indonesian positive law provides a normative framework to address these challenges, although its application remains contested.
Law Number 2 of 2004 establishes procedural mechanisms for resolving industrial relations disputes, while Law Number 27 of 2022 on Personal Data Protection grants data subjects the right to object to decisions based solely on automated processing that produce significant legal effects.
In addition, the Law on 438 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 Information and Electronic Transactions recognizes electronic information and documents as valid legal evidence (Law No.
2 of 2004 Concerning the Settlement of Industrial Relations Disputes.
Law Number 27 of 2022 Concerning Personal Data Protection.
Article 10.
Article 32, and Article 34.
Law Number 11 of 2008 Concerning Information and Electronic Transactions as Last Amended by Law Number 1 of 2024.
Article 5, n.
Furthermore, the Constitutional Court Decision Number 168/PUU-XXI/2023 reinforces that termination of employment cannot be unilaterally effected without due legal process, underscoring the necessity of institutional oversight in employment decisions.
The legal framework shows that work decisions based on automated systems can actually still be tested normatively.
However, the main challenge lies in how to read the case.
Employers often position system results as objective data, while workers often struggle to access basic data, processing track records, and substantive reasons used to justify company Therefore, without a critical approach, industrial relations dispute forums risk only examining the formality of the decree without penetrating the digital mechanisms that give birth to justice for the workers themselves.
The research gap is evident in the existing literature, which has predominantly focused on the technological and efficiency dimensions of algorithmic management, while insufficiently examining its concrete legal implications within the Indonesian industrial relations framework, especially in relation to reassignment practices based on automated systems (Irma Cahyaningtyas and RoAofah Setiawati 2.
Recent legal developments also show that termination of employment should not be allowed to be a unilateral action that can be easily wrapped up in any managerial terms.
In the Constitutional Court Decision Number 168/PUU-XXI/2023, the Court emphasized that if bipartite negotiations do not reach an agreement, then termination of employment can only be carried out after obtaining a determination from the industrial relations dispute settlement institution whose decision has permanent legal force.
This decision is relevant to be read together with the issue of automatic decisions, because he emphasized that the termination of the employment relationship should not be disguised or accelerated only based on the company's internal instruments (Mahkamah Konstitusi Republik Indonesia.
Putusan Nomor 168/PUU-XXI/2023 tentang Pengujian Undang-Undang Nomor 6 Tahun 2023 terhadap Undang-Undang Dasar Negara Republik Indonesia Tahun 1.
The main problem addressed in this study lies in the potential for automated decisionmaking systems to produce substantively unjust outcomes in industrial relations, particularly when used to justify worker reassignment policies that disadvantage employees.
The urgency of this research stems from the increasing adoption of digital management systems in employment practices, which has not been adequately accompanied by corresponding legal safeguards, thereby risking the erosion of workersAo rights and protections.
Based on this background, the problems studied in this paper focus on two things.
First, what is the legal position of the work decisions produced by automated systems in industrial relations in Indonesia.
Second, how to resolve industrial relations disputes over these decisions, especially in the case of moving workers to branches far from their residences to encourage resignation and avoid severance pay obligations by the Company With this focus, this research is expected to show that the digitalization of employment relations should not be understood solely as a technological innovation, but should be read as a change in governance that demands responsibility New and stricter laws.
METHOD
This research is a normative legal research.
The choice of method is based on the character of the problem being studied, namely issues regarding legal norms, legal principles, and the construction of legal accountability for work decisions produced by automated Normative legal research is used to examine how the applicable law places 439 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 automatic decisions in industrial relations, how the law provides protection to workers, and how the dispute resolution mechanism is carried out in accordance with relevant laws and regulations (Marzuki 2.
The approaches used in this study include a legislative approach, a conceptual approach, and a case approach.
The legislative approach is carried out by examining the Constitution of the Republic of Indonesia in 1945.
Law Number 2 of 2004 concerning the Settlement of Industrial Relations Disputes.
Law Number 27 of 2022 concerning the Protection of Personal Data.
Law Number 11 of 2008 concerning Information and Electronic Transactions as last amended by Law Number 1 of 2024, and Constitutional Court Decision Number 168/PUU-XXI/2023.
A conceptual approach is used to develop analysis through the theory of legal protection, legal certainty, procedural justice, and algorithmic accountability.
Meanwhile, the case approach is used to read the illustration of mutations to distant branches as a form of covert pressure on workers.
The primary legal materials in this study are in the form of relevant laws and regulations and court decisions, while secondary legal materials are in the form of books, journal articles, and institutional publications that discuss industrial relations, personal data protection, electronic proof, and algorithmic management in the workplace.
The technique of collecting legal materials is carried out through literature studies.
All legal materials are then analyzed qualitatively by connecting the applicable norms with legal facts that may arise in the use of automated systems in the work environment.
The analysis method used is descriptive-analytical, which is to outline relevant norms and concepts, then assess how these norms should be applied to protect workers when work decisions are generated by an automated system.
Through this method, this study not only aims to photograph the existing legal situation, but also assess the sufficiency of Indonesia's positive law in dealing with new forms of industrial relations disputes born due to the digitalization of work management.
Thus, this paper is expected to make a conceptual contribution to the reading of labor law in the era of data-based decision- making and automated systems.
RESULTS AND DISCUSSION
Legal Status of Work Decisions Generated by Automated Systems in Industrial Relations In Indonesian industrial relations law, the analytical starting point must be grounded in the identification of legally recognized subjects.
Law Number 2 of 2004 defines industrial relations disputes as conflicts between employers .
r their association.
and workers/labor unions, thereby affirming that legal relations exist exclusively among human actors or legal entities acting through their organs, not between workers and technological systems (UndangUndang Nomor 2 Tahun 2004 Tentang Penyelesaian Perselisihan Hubungan Industrial 2.
Consequently, automated systems, regardless of their sophistication, do not constitute independent legal subjects and cannot bear legal responsibility.
This construction leads to a crucial implication: all employment decisions generated through automated systems must be legally attributed to the employer.
Digital systems function merely as instruments for data processing, scoring, and recommendation-making.
The legal consequences arise not from the system itself, but from the employerAos decision to adopt and operationalize its outputs.
Accordingly, liability for any adverse impacts including loss of rights, unfair reassignment, or termination, remains fully attached to the employer as the controlling and benefiting party.
However, this formal attribution of responsibility becomes problematic when examined through a critical lens of algorithmic bias.
Automated systems are inherently dependent on historical data and predefined parameters, which may embed structural biases against certain 440 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 categories of workers.
For instance, performance metrics that prioritize constant availability, uninterrupted attendance, or output maximization may disproportionately disadvantage workers with caregiving responsibilities, health conditions, or limited access to digital tools.
Such biases are often invisible at the surface level yet systematically reproduced through automated evaluation processes.
Therefore, the assumption that algorithmic decisions are inherently objective must be rejected, as they may instead amplify pre-existing inequalities within the workplace.
In this context, the theory of legal protection articulated by Philipus M.
Hadjon becomes highly relevant.
Legal protection entails safeguarding human dignity and ensuring that legal subjects are not subjected to arbitrary treatment.
Within algorithmically mediated employment relations, this principle requires more than formal acknowledgment of rights.
demands substantive access to the basis of automated decisions, including the ability to identify and challenge biased or inaccurate data inputs.
Without such access, workers are effectively positioned as passive objects of opaque decision-making systems (Philipus M.
Hadjon 1.
The issue of transparency is further reinforced by Law Number 27 of 2022 on Personal Data Protection, which grants data subjects the right to obtain information regarding data processing purposes, legal bases, and accountability of data controllers, as well as access to their personal data and its processing history In employment contexts, where performance assessments are heavily data-driven, these rights should be interpreted expansively to include transparency of evaluation criteria, scoring mechanisms, and decision thresholds (Law Number 27 of 2022 Concerning Personal Data Protection.
Article 4.
Article 6.
Article 32, and Article 34 2.
More critically.
Article 10 of the same law provides the right to object to decisions based solely on automated processing that produce legal or significant effects This provision introduces an implicit requirement of human intervention and review in employment However, in practice, companies often treat system outputs as final determinations, thereby undermining the normative intent of the law.
This reveals a gap between formal legal guarantees and their implementation in digital workplace governance (Law Number 27 of 2022 Concerning Personal Data Protection.
Article .
From the perspective of legal certainty, automated decision-making systems must satisfy the principles of clarity and predictability.
Workers cannot reasonably be expected to comply with performance standards or behavioral expectations that are neither disclosed nor If algorithmic systems determine key employment outcomes, then their underlying parameters must be explicitly incorporated into employment contracts, company regulations, or collective agreements.
Otherwise, such systems generate normative uncertainty, as workers are subjected to evaluative standards that are effectively hidden.
Equally important is the principle of procedural justice, which emphasizes that fairness is determined not only by outcomes but also by the integrity of the decision-making process.
In the context of automated systems, procedural justice requires the preservation of core due process elements: prior notification, the right to be heard, the opportunity to provide explanations, and access to effective remedies.
Automated notifications of reassignment, sanctions, or termination without meaningful explanation or contestation mechanisms fail to meet these standards and risk violating fundamental fairness principles (E.
Allan Lind 1.
This concern also intersects with the concept of Pancasila industrial relations, which prioritizes deliberation, balance, and respect for human dignity.
Algorithmic management, if left unchecked, risks reducing workers to quantifiable data points, thereby eroding the human-centered foundation of labor law.
Consequently, human oversight should be regarded not as an optional feature but as a minimum legal requirement in decisions with significant 441 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 At the level of corporate responsibility, the use of automated systems raises questions of accountability for algorithmic governance.
Employers cannot evade liability by invoking technological objectivity or system autonomy.
Instead, they bear a duty to ensure that the systems they deploy are transparent, non-discriminatory, auditable, and aligned with legal This includes obligations to regularly evaluate system performance, correct biased outputs, and provide accessible mechanisms for worker.
Failure to do so may constitute not only a breach of labor law but also a violation of data protection obligations.
International discourse supports this position.
The International Labour Organization highlights that algorithmic management may intensify surveillance, obscure evaluation criteria, and reduce worker participation, thereby increasing the risk of injustice.
Similarly, the OECD identifies transparency and accountability as central challenges in the deployment of algorithmic tools in the workplace.
These findings reinforce the need to subject automated systems to rigorous legal scrutiny rather than treating them as neutral managerial tools.
(Annarosa Pesole et al.
In evidentiary terms, the ITE Law recognizes electronic information and documents as valid legal evidence, including system logs, performance records, and digital However, such evidence should not be conflated with substantive truth.
Electronic records merely document the operation of a system.
they do not automatically validate the fairness or legality of the resulting decisions.
Judicial and quasi-judicial bodies must therefore critically assess not only the existence of digital evidence but also the integrity of the processes that produced it (Law Number 11 of 2008 Concerning Information and Electronic Transactions as Last Amended by Law Number 1 of 2024.
Article 5 2.
Based on this analysis, the legal position of automated employment decisions in Indonesian law can be reformulated in a more critical and structured manner.
First, automated systems remain auxiliary instruments and do not possess legal subjectivity.
Second, all legal consequences arising from system-based decisions are attributable to the employer.
Third, workers possess enforceable rights to transparency, access, and objection against automated decisions that significantly affect them.
Fourth, the legality of such decisions must be assessed not only in terms of formal compliance but also through substantive criteria, including data integrity, absence of bias, transparency of parameters, and procedural fairness.
With this construction, it becomes evident that Indonesian law provides a normative foundation to challenge the assumption that automated decisions are inherently objective or legally authoritative.
Instead, such decisions must remain open to scrutiny, contestation, and correction within the broader framework of labor protection and justice.
Table 1.
Automated System-Based Work Decision Mapping and Its Dispute Implications Form of Decision Example Practice Object of Dispute Focus of Legal Tests Target, compliance, or Indicator transparency.
Performance Disputes of rights or productivity scores are data accuracy, and lowered by the system correction rights Legal basis for Alerts or restrictions on Digital sanctions work access based on Disputes of rights proportionality, and automated notifications defense of workers Move to a distant Disputes of rights.
Operational needs, branch after Automatic mutation interests, or disguised policy motives, and performance score humanitarian impacts Deactivation or Termination of Validity of the reasons termination of work Termination of for layoffs, procedures, based on system employment dispute and post-layoff rights 442 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 The table above shows that a single automated decision can give rise to more than one type of dispute.
Therefore, the reading of the object of dispute must always depart from the real consequences experienced by the worker, not solely from the administrative label used by the employer.
Industrial Relations Dispute Resolution on Automated System Decisions and Case Analysis of Job Transfers to Distant Branches Within the framework of Law Number 2 of 2004, automated system-based work decisions can give birth to several types of disputes.
When the system causes workers' financial rights to be reduced, for example, wages, incentives, or bonuses are not paid because workers are considered not to meet certain indicators, then the problem can fall into the category of rights disputes.
When a company uses a system to change work conditions, targets, or placement patterns without a legitimate basis, its disputes can lead to conflicts of Meanwhile, when the system is used as an excuse to terminate an employment relationship, either directly or covertly, the dispute can develop into a termination dispute (Law Number 2 of 2004 Concerning the Settlement of Industrial Relations Disputes.
Article 1 Number 2.
Number 3, and Number 4 2.
Mapping this type of dispute is important because the evidentiary strategies, forms of demands, and legal argumentation paths can differ.
In the case of an automated decision, a single set of actions can give birth to more than one dispute.
For example, a decrease in performance score can lead to a reduction in bonuses or incentives first.
After that, the same score is used to move workers to a very distant branch.
In the next stage, when the worker objected and was unable to accept the mutation, the company considered the worker to be absent or disloyal and pushed him to resign.
The series shows that rights disputes, interest disputes, and layoff disputes can be intertwined in one automated system-based employment relations event.
The first stage of dispute resolution remains mandatory through bipartite negotiations.
The law affirms that industrial relations disputes must first be resolved through deliberation to reach a consensus within a maximum of thirty working days from the start of negotiations.
If one of the parties refuses to negotiate or after that period of time no agreement is reached, the bipartite is considered a failure.
This principle is important because the bipartite forum is the starting point to test the company's good faith and provide an opportunity for workers to request data disclosure, explanation of assessment indicators, and reasons for using the results of the system as a basis for policy (Law Number 2 of 2004 Concerning Industrial Relations Dispute Resolution.
Article 3 2.
If the bipartite fails, the dispute is recorded with the agency responsible for the field of After registration, the parties are offered the option of conciliation or If there is no agreement, the settlement is delegated to the mediator.
Research on industrial relations dispute resolution confirms that mediation has a central position because it is able to bring the parties together and be a means of balancing before the dispute is brought to court.
In cases stemming from automated decisions, the mediator should play a more active role in examining not only the company's decision, but also the data, logs, processing track record, and rationality of the parameters used by the system (Ana Silviana 2.
In principle, the role of mediators in automated system-based disputes should not be purely administrative.
The mediator needs to assess whether the worker is really informed about the parameters of the system, whether there is access to the correction mechanism, whether the company has ever conducted an audit of the accuracy of the data, and whether the decision taken is still proportionate to the company's operational objectives.
If these questions are ignored, mediation will only repeat the logic of the system without placing workers as subjects who have the right to be heard.
443 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 In the event that the dispute continues to the Industrial Relations Court, electronic evidence becomes very important.
Article 5 of the ITE Law regime affirms that electronic information, electronic documents, and printed results are legal evidence.
Therefore, performance assessment dashboards, biometric attendance records.
GPS history, notification logs, e-mails, and printouts from the system can be submitted as evidence.
However, the judge should not stop at the formal acceptance of electronic documents.
The judge must still test the validity of the data source, the integrity of the digital footprint, the relationship of the data with the policies taken, and the possibility of bias or errors that are detrimental to workers (Law Number 11 of 2008 Concerning Information and Electronic Transactions as Last Amended by Law Number 1 of 2024.
Article 5 2.
Especially in cases related to termination of employment, the Constitutional Court Decision Number 168/PUU-XXI/2023 provides a very important protection direction.
The Court interpreted that if bipartite negotiations do not reach an agreement, then termination of employment can only be carried out after obtaining a determination from the industrial relations dispute settlement institution whose decision has permanent legal force.
The Court also emphasized that the dispute resolution process must be understood until it ends with a decision that has permanent legal force.
This formulation narrows the space for employers to carry out unilateral layoffs, either openly or through covert schemes (Constitutional Court of the Republic of Indonesia.
Decision Number 168/PUU-XXI/2023 Concerning the Examination of Law Number 6 of 2023 against the 1945 Constitution of the Republic of Indonesia 2.
The case of moving workers to branches far from where they live needs to be read in that framework.
Formally, the company may refer to it as a mutation.
However, the law should not stop at formal labels.
Labor law studies show that mutations are often abused as a pressure tool, especially when they are carried out to distant locations, without a clear operational reason, or followed by a reduction in positions and burdensome working In such circumstances, mutation is no longer neutral as a managerial policy, but has the potential to be a form of abuse of authority to force workers out of the company (Angling Dandy Wicaksana 2.
If the mutation to a distant branch is carried out after the system assesses that the performance of workers has decreased, then there are several layers of legal problems.
First, the validity of the performance score itself must be tested.
Workers have the right to know what data the system uses, how it is processed, whether there are input errors, and whether contextual factors such as equipment interference, unbalanced workload, or workers' health conditions are taken into account.
Second, the operational reason behind the mutation must be proven.
Companies must be able to demonstrate that placement to a distant branch is indeed based on real organizational needs, not simply because workers are seen as problematic or want to be set aside.
Third, the proportionality of the mutation policy must be Even if the operational need exists, mutations that actually create an excessive burden and are not accompanied by reasonable support can still be seen as contrary to the principle of propriety.
Here, the right to object to automated decisions as provided for in the Personal Data Protection Act acquires practical relevance.
Workers can argue that the mutation or decline in their work status was born from a decision that was heavily influenced by automated processing and had a significant impact on their lives.
On that basis, workers can request access to data and processing tracks, demand correction of inaccurate data, and reject decisions taken without meaningful human scrutiny.
This kind of construction confirms that personal data law can be used in a complementary way with industrial relations law to strengthen worker protection (Undang-Undang No.
27 Tahun 2022 Tentang Perlindungan Data Pribadi Pasal 10, 13, 32.
Dan 34 2.
444 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 From the point of view of substance, resignations born of such mutation pressures cannot be automatically treated as fully autonomous free will.
The law must assess the reality of the power relations that shape the decision.
When workers are faced with the choice between accepting an inhumane placement or leaving a job without the protection equivalent of layoffs, then the element of structural pressure becomes very strong.
In such circumstances, the resignation can be viewed as the result of a company policy that materially resembles the termination of a covert employment relationship.
This analysis is in line with research on the rights of workers who resign, which confirms that workers who actually resign of their own volition do not in principle receive severance pay, but only certain rights such as compensation money and separation money according to applicable regulations.
Therefore, employers have an economic incentive to wrap the termination of employment as voluntary resignation.
This is where the industrial relations dispute forum is obliged to penetrate the formal form and read whether the resignation is really born from free will or precisely from the pressure of corporate policy (Intan Rakhmawati 2.
Thus, the settlement of industrial relations disputes over work decisions produced by automated systems must be based on the principle of substantive examination.
Bipartite forums, mediation, and Industrial Relations Courts should not only assess the existence or absence of a letter of mutation, the existence or absence of a performance score, and the existence or absence of a letter of resignation.
What must be tested is the entire chain of decisions: whether the underlying data is correct, whether the assessment mechanism is reasonable, whether the mutation is proportional, whether the worker is given room to defend himself, and whether the company's overall actions are actually directed at avoiding severance pay.
If proven to be so, then the automated system has been used not as a tool of efficiency, but as an instrument of legal smuggling that should be corrected by the mechanism of industrial relations disputes.
In the end, the digitization of employment relations requires a change in the perspective of employment law.
Employers cannot take refuge behind the phrase "system results" to deny Instead, dispute resolution officers must become increasingly accustomed to assessing data, processing logic, and electronic evidence while maintaining the principles of fairness, balance, and protection for weaker parties.
As long as automated systems are used to make decisions that have a significant impact on workers, the law must always ensure one thing: that those decisions remain questionable, examined, and accountable by the people behind them (Ira Hanifah 2.
CONCLUSION
Main Findings This study finds that employment decisions generated by automated systems in Indonesian industrial relations law remain legally attributable to the employer.
Automated systems, including algorithms and digital applications, do not possess legal subjectivity and therefore cannot bear independent responsibility.
Consequently, all legal consequences arising from such decisions are attached to the employer as the party who designs, controls, and implements the system.
Automated decision-making must therefore be subject to legal scrutiny based on fundamental criteria, including the legality of its basis, the quality and accuracy of data, the transparency of parameters, and the fulfillment of procedural fairness.
Workers retain the right to access, assess, and object to decisions, particularly when such decisions are based solely on automated processing and produce significant legal or practical impacts.
In dispute resolution, automated system-based decisions must be examined through existing industrial relations mechanisms, including bipartite negotiations, mediation, and 445 | P a g e https://research.
e-greenation.
org/GIJLSS, size .
Vol.
No.
May Ae June 2026 adjudication before the Industrial Relations Court.
Importantly, legal assessment should not be limited to the formal justification of company policies, but must extend to their substantive intent, proportionality, and potential use as covert pressureAisuch as in cases of reassignment to distant locations aimed at inducing resignation and avoiding severance obligations.
Thus, the increasing use of digital systems does not diminish legal protection for workers but instead heightens the need for accountability.
Regulatory Recommendations Employers should position automated systems strictly as decision-support tools rather than as autonomous determinants of employment outcomes.
The use of such systems must be transparently regulated within employment agreements, company regulations, or collective bargaining agreements, including clear performance indicators, data verification mechanisms, objection procedures, and mandatory human oversight prior to the implementation of decisions with significant impact.
At the institutional level, mediators, conciliators, and industrial relations judges must enhance their capacity to assess electronic evidence, algorithmic processes, and issues of procedural fairness in digital decision-making contexts.
Furthermore, the state should develop more specific regulatory frameworks governing algorithmic accountability in employment, ensuring that the deployment of technology aligns with the principles of legal certainty, social justice, and the protection of workersAo dignity.
REFERENCES