The 2026 International Labour Conference (ILC) adopted a hard-fought and historic new international convention — a binding international labor standard for the platform economy. Convention No. 193 on Decent Work in the Platform Economy was adopted after two years of tripartite negotiation between governments, employers, and unions. The Convention establishes that fundamental protections — occupational safety and health, social protection, non-discrimination, freedom of association and collective bargaining — apply to platform workers regardless of their employment classification. It requires governments and employers to deal with the widespread misclassification of workers as contractors and to facilitate formalization. Furthermore, the Convention addresses algorithmic management in a binding international instrument for the first time – potentially advancing the case for digital justice.
The Convention was not universally supported; the governments of the United States and New Zealand voted against the Convention, and several governments and employers abstained.
The Convention was not universally supported; the governments of the United States and New Zealand voted against the Convention, and several governments and employers abstained. The employers’ spokesperson routinely opposed or sought to amend articles that placed a positive obligation on governments or employers to protect rights. Instead, they sought to have principles applied “according to national law and practice” (or continue doing as you see fit). Yet, the majority of governments, employers and, of course, workers, finally voted in favor of the amended instrument. The sustained delays during the negotiations, however, resulted in a failure to adopt the important and complementary Resolution, which is designed to provide further guidance on how to implement the obligations.
We shouldn’t need C193
It’s shameful that this Convention was needed. Every worker has a right to fundamental labor rights and no worker, regardless of the way their work is allocated, should have been denied those rights. But digital labor platforms have spent the past decade building business models designed to sever the legal relationship between capital and labor, while keeping every operational lever of an employer — task allocation, pricing, discipline, deactivation — firmly in Big Tech’s hands. As I reported to PSI affiliates following the vote, this is not a side effect of new technology; it’s a result of unregulated corporate power. Platform corporations’ “innovations” were to find ways around labor standards; they have “misclassified” workers as independent contractors to strip them of the rights that decades of union struggle secured.
Every worker has a right to fundamental labor rights and no worker, regardless of the way their work is allocated, should have been denied those rights.
The consequence has been the dismantling of the social contract. Workers allocated work by automated systems frequently have no employment contract, no social protection, and no enforceable right to organize. Meanwhile, the same systems that allocate their work also harvest granular data about how they perform it. Location, speed, acceptance rates, customer ratings, biometric signals (this could include heart rates, eye movements, facial expressions) are all data points that platforms use to optimize profits while workers are told almost nothing about how decisions affecting their payments or allocation of work are made.
This is not confined to ride-hailing and food delivery. Platform work is expanding into the delivery of public services, above all in health and care. Home-based carers, disability support workers, aged care workers, nurses and allied health professionals are increasingly booked, tracked, rated and paid through digital platforms that fragment continuity of care into discrete, billable tasks. These are overwhelmingly women workers, often migrants, already underpaid and undervalued. As algorithmic scheduling spreads further into care systems, the same dynamics of surveillance and unaccountable automated decision-making are appearing — with consequences for both workers and the people who need care. Consequently, PSI played an active role in advocating for a comprehensive convention and recommendation that would address the exploitative nature of platformization.
Strengths of the Convention
The Convention applies to all workers “engaged or employed to work for the provision of service organized and/or facilitated by a digital labour platform” regardless of whether they have been recognized as an employee or not. Importantly, obligations also apply to intermediaries who make platform work available. The obligation to provide at least minimum wage (Article 10.2), however, is restricted to employees, with states encouraged to ‘consider’ ensuring self-employed workers are also covered.
The obligation to provide at least minimum wage (Article 10.2), however, is restricted to employees, with states encouraged to ‘consider’ ensuring self-employed workers are also covered.
To deal with the systemic failure to recognize platform workers as employees, Article 9 of the convention requires member states to “take appropriate measures to ensure the correct classification of digital platform workers” and, in doing so, to be “guided mainly by the facts relating to the performance of work ”. Workers sought to have the Employment Relationship Recommendation (R198) referenced in the Convention to spell out some of the possible ‘facts’ that should be used to determine whether a worker is an employee, including “the fact that the work: is carried out according to the instructions and under the control of another party” (Article 13a of R198).
While the reference to the Recommendation was not ultimately included, these facts should still guide the determination of classification, and there is still scope to include a reference to R198 in the future recommendation.
The Convention also deals with occupational safety and health, violence and harassment, remuneration, social security, workers’ data privacy, deactivation of accounts, terms and conditions, protection of migrant workers, and dispute resolution and remedies.
Algorithmic accountability and explainability
Amongst the most contested elements in the draft convention were workers’ proposals to have access to information about the use, intended purpose and consequences of data collected and the algorithmic logic. The 2025 negotiations included a proposal that workers and their representatives should have access to the source code used to generate algorithms. US employers were so incensed with the suggestion that they urged the US government to stay in the ILO just to oppose it.
The right of workers and their unions to understand, contest and act on the automated decisions shaping their working lives should be fundamental.
The right of workers and their unions to understand, contest and act on the automated decisions shaping their working lives should be fundamental. Many other rights are contingent on this right because, without it, unions cannot identify underpayments, contest discrimination, assess occupational health and safety threats or engage in collective bargaining. Unions have fought for the right to information about work environments and working conditions for more than a century. When the first Factory Laws were proposed in nineteenth-century Britain, unionists and those opposed to child labor argued that mills and factories needed to be inspected to verify that children were not working fourteen-hour shifts in lethal conditions. Then, factory owners argued that the factory floor was private property and commercially sensitive. Asbestos companies argued the composition of insulation products and the health studies they had were proprietary, commercial secrets. There are many other examples of how unions have fought, and together these cases lead to the occupational safety and health (OSH) standards we have achieved. That’s why in some countries OSH laws are termed ‘Workers’ Right to Know’ laws (including in the US). And it’s why Convention 155 on Occupational Safety and Health says that (Article 19) Workers’ safety delegates …(and) other workers’ representatives should be given adequate information on safety and health matters, enabled to examine factors affecting safety and health.
Despite this, it took significant negotiation and some compromise to finally achieve section X in the convention on ‘Impact of the use of automated systems’ — containing three valuable articles. Together, these articles establish the first international right for workers and their representatives to be informed of any automated systems used to monitor, evaluate or make decisions about work.
Article 13 reads:
Each Member shall require digital labour platforms to inform digital platform workers, before their employment or engagement, and their representatives or representative workers’ organizations and, where they exist, organizations representing digital platform workers, about:
- the use of automated systems, based on algorithms or on similar methods, to monitor or evaluate work, or to generate decisions relating to work;
- the extent to which the use of such automated systems has an impact on the working conditions of digital platform workers or their access to work.
Article 14 restricts platforms from using automated systems in ways that conflict with fundamental labor rights by requiring states to ensure ‘responsible use’ consistent with obligations to realize the fundamental principles and rights at work — meaning algorithms cannot be used in ways that expose workers to occupational health and safety risks (for example, enticements for delivery drivers to speed), discrimination (i.e, paying migrant workers or mothers less because they are more likely to take low payment jobs), or undermine their trade union rights.
Workers also have the right to receive a written explanation of decisions resulting in non-payment, suspension, deactivation, or termination, and to ensure “human involvement” in a review of adverse decisions (Article 15).
The parameters of the information workers and their unions must receive should be further detailed in the Recommendation. Ideally, the Recommendation should outline the elements of explainability. It needs to detail comprehensive, accessible information about a system’s purpose, design, functioning and logic before it is used on them, and after any update; and a right to written explanations of specific automated decisions, including the actual parameters and values considered, in language workers can understand. These are the building blocks of meaningful accountability — turning the black box of algorithms into something a union representative can actually interrogate, challenge and bargain over.
These are the building blocks of meaningful accountability — turning the black box of algorithms into something a union representative can actually interrogate, challenge and bargain over.
Similarly, the Recommendation needs to outline the role of labor inspectorates in interrogating the impact of automated systems and ensuring compliance. Labor inspectorates need to have access to similar parameters of data and build capacities to interpret algorithms and their impact.
Some jurisdictions have already recognized the need for workers and their unions to have access to algorithmic data and the logic behind automated decisions. For example, in Spain, the Riders’ Law was enacted partly because Spanish labor courts found that effective collective bargaining was impossible without algorithmic transparency. The law requires employers to inform workers’ representatives, or the works council, of the parameters, rules and instructions on which the algorithms or AI systems are based.
The law also expands rights of entry for union representatives designated as health and safety representatives who are entitled to reasonable assistance to access and inspect digital work systems where a breach of Work Health and Safety (WHS) laws is suspected.
In Australia, the state of New South Wales amended its work health and safety laws to expressly cover risks created by “digital work systems”, including psychosocial risks. The law also expands rights of entry for union representatives designated as health and safety representatives who are entitled to reasonable assistance to access and inspect digital work systems where a breach of Work Health and Safety (WHS) laws is suspected. And in this case, the law applies to all workers, not only those engaged through digital platforms. This is important as automated management, algorithmic scheduling, and data-driven monitoring tools are not unique to platforms; they are being absorbed into conventional employment relationships, and a future objective for the union movement will be to ensure all workers enjoy this level of protection.
What comes next
Convention 193 will come into force 12 months after two governments ratify it. Unions have already begun campaigning for ratification and using the instrument to push for national changes, even without ratification. But the reality is that unions will also need to develop much more capacity to utilize the rights won. In Spain, the law has been under-utilized because “neither unions nor the labor inspectorate are equipped to handle the complexities of algorithmic tools”. In the relentless global attack on labor rights, trade unions and public services, there are too many urgent demands on under-resourced unions to utilize hard-fought reforms. It’s therefore essential that digital justice movements, who have the capacity to scrutinize algorithmic management systems, work with the union movement to challenge Big Tech and expose their efforts to undermine the social contract.
