What will the new Convention on Decent Work in the Platform Economy mean for workers and what's next?

We explored what the recently adopted ILO Convention on Decent Work in the Platform Economy will mean for workers and labour rights, and analysed what are gaps that still need to be addressed during and beyond its implementation into national legislation.

Key points
  • The Convention is a major step forward for workers' rights in the digital economy.
  • Some of the Convention's most innovative provisions relate to the use of automated systems, and the need to protect workers from their adverse effects.
  • The Convention text itself is not perfect as some language around automated decision-making (ADM) defaults to vague terms.
  • Stronger safeguards for workers against the harms of automated decision-making are still required.
Advocacy
Gig worker on motorcycle being tracked.

On 12 June 2026, the International Labour Conference adopted the C193 Convention on Decent Work in the Platform Economy. Overall, this is a landmark piece of legislation, as the first global legally binding instrument regulating the platform economy and addressing the impact of technology on digital platform workers, in particular in relation to the use of automated-decision making. Once a government ratifies the Convention, it will be legally obliged to enforce its innovative provisions, including in relation to the use of automated-decision making at work. 

A major step forward for workers' rights in the digital economy

This Convention represents a significant step forward for workers' rights. It reflects the growing consensus that the rights of digital platform workers need protecting from abusive practices of digital platforms, and on the need to regulate the use of automated-decision making in the workplace.

The treaty codifies the reality that platform workers are indeed workers - rather than merely contractors with platforms as the intermediaries. The result is that the ILO framework of Fundamental Principles and Rights at Work applies, along with new sector-specific rules, meaning that workers have rights, and platforms have duties towards them. Convention C193 covers a wide range of issues that must be addressed to prevent platform companies being able to exploit workers and their data through the use of technology in the workplace. While not all measures extend equally to all workers, it contains important provisions regarding: fairer compensation commitments for platform workers; the extension of access to certain social protection rights for gig workers; and requiring governments to ensure workers are classified more appropriately for their work.

The treaty also acknowledges that certain rights apply regardless of employment classification, including the right to collective bargaining and protection from forced or child labour. Meaningful positives were also given to workers with regards to occupational health and safety and less randomised termination.

The Convention crucially recognises the need for greater transparency in algorithmic management, requiring companies to inform workers about automated systems used to monitor performance, allocate tasks and make decisions affecting their work and access to work. Workers can also request written explanations of significant automated decisions, as well as request review of decisions that negatively impact them.

The final text also introduces some valuable additions on workers' rights to request access, erasure and rectification of their personal data processed by digital labour platforms.

Both of these represent milestones as the first time that a global convention has set out binding legal requirements for how digital technology is used to manage workers, and what rights people have over the data on them collected when at work.

Ratification and then enforcement now lies with the individual Member States, where we will see how the Convention may practically come to life for workers -- and whether it's effectively interpreted and leads to the adoption of national legislation which protects workers and holds to account powerful actors in the platform economy.

Automated systems and workers' rights

Our focus at PI is on the need to protect digital workers against the negative effects of the use of automated systems and the abuse of personal data associated with it. These were central to our advocacy efforts throughout the drafting of the ILO Convention. With some of the Convention's most innovative provisions on these topics, we hope they will shape international labour laws moving forward.

Section X of the Convention requires companies to inform workers about automated systems used to monitor or evaluate their work. It also gives workers the right to receive a written explanation of significant automated decisions, and the power to request a review (with appropriate human involvement) of certain decisions , including nonpayment, suspension, deactivation or termination of their employment. Under this section, the Convention imposes a welcome obligation on governments to ensure that automated systems are used responsibly and with regards to fundamental principles and rights at work.

We also welcome that Section XI requires Member States to establish effective and appropriate safeguards concerning digital platform workers’ personal data. More specifically, Article 16 asserts that personal data collected in the course of platform work may only be processed for the legitimate purpose for which it was gathered, and enshrines the right of workers to request access to, and the rectification and erasure of, their personal data processed by digital labour platforms. This is a crucial protection for workers who are subjected to significant levels of data extraction and surveillance in the performance of their work.

The next step is ensuring Member States adequately enforce these provisions in workers' best interests, as well as investigating the need for further avenues of protection in addition to these historical gains in the Convention.

Gaps to watch out for

The Convention text itself is not perfect, as some language around automated decision-making (ADM) defaults to vague terms. For instance:

  • Rather than hard-edged requirements, Member States are only obliged to 'take appropriate measures' to ensure responsible use of automated systems (Article 14) and access to explanations and reviews for automated decisions (Article 15).
  • Worker access to a written explanation pertains only to 'significant' decisions, without defining what constitutes a significant decision (Article 15(1)(a)).
  • Reviews of decisions, and the degree of human involvement in them, must be 'appropriate' - but again there is no definition of this term (Articles 15(1)(b) and 15(2)).
  • The text limits data processing only for a 'legitimate purpose' but it does not define nor qualify what this would be (Article 16).
  • The text limits the information to be provided to workers in relation to the terms and conditions of their employment to what is 'in accordance with national law, regulation and practice' (Article 18).
  • The text requires states to take appropriate measures to 'protect commercially sensitive information', which may be used to undermine transparency obligations (Article 24(4)).

Open-ended language like 'appropriate measures' leaves too much to discretion, rather than setting a clear standard of what is expected to ensure the essence and purpose of the Convention is respected. The ambiguous language and lack of clarity leaves the door open for Member States (and platforms) to take an approach of implementing the bare minimum, therefore reducing accountability and ultimately failing to protect workers. As States approach national implementation, they must ensure that they clearly articulate the nature of those 'appropriate measures' in a bold and ambitious way. Setting the threshold at which protections kick in at an appropriate level requires challenging, not reinforcing, the power imbalance against workers.

Similarly, the strength of Article 16 regarding regulating the processing of workers' personal data could hinge on how Member States define 'legitimate purpose' in domestic implementation, as data collected for managing labour could still be repurposed through more expansive definitions. An overly narrow approach as to whether data constitutes personal data may also undermine the impact of Article 16 - for example if platforms seek to argue that aggregate datasets or pseudonymised data should be excluded.

Much of the Convention's substantive language is also qualified by phrases such as 'in accordance with national law, regulation and practice', for example Article 18, which makes the underlying Convention guarantees only as strong as what the Member States choose to interpret as per existing national laws or implement into new national legislation. The purpose of an international agreement is to agree new common standards and raise the bar: this cannot be held hostage to existing rules which are now known to be in need of improvement.

We also cast some caution over the Convention's provisions protecting platforms' commercially sensitive information under Article 24 (4). This may have the effect of undermining how much system logic actually has to be disclosed in practice. Platforms may seek to weigh algorithmic transparency as in conflict with proprietary secrecy over trade secrets. As a result, how Member States resolve these definitional gaps will determine whether transparency rights give workers substantial insight into automated systems or are held at bay by broad claims of 'commercial secrecy'.

Interpreting and implementing

As Member States approach national implementation, it is crucial that they, in consultation with workers' representatives, civil societies and independent technical experts, look beyond the Convention towards best practices and embed provisions to: define what constitutes 'appropriate measures' to protect workers' rights from being negatively affected by the use of ADM; not excessively limit explanations only to a very narrow range of decisions; and clarify 'appropriate human involvement' such that human oversight has measurable impact in the quality of decisions and the reality of workers' experience.

Given the growth in problematic practices observed in the platform economy, measures may need to be transformative in order to be considered 'appropriate'. For instance, appropriate measures might mean that platforms are required to maintain a public register that explains the purpose, functioning and capabilities of automated decision-making systems or algorithms used to manage workers. They may also need to provide workers and their representatives with access to sandboxes, source code and technical expertise to adequately understand and challenge those decisions. The existing power asymmetry where platforms themselves have the most visibility into the algorithms while workers are operating from a knowledge base of far less, can weaken workers' efforts to challenge decisions.

Member States should also push platforms to provide explanations of decisions to workers by default, instead of just 'upon request' and any decisions that determine a workers' job environment (e.g., job allocation, pay, account deactivation, social protections) should be considered significant. Explanations provided to workers should specify the parameters and inputs used to make the decisions, and the explanations should be specific and personalised to the worker rather than wholly generic or overly technical. For example, the explanation should state that ‘on this date you were expected to make X deliveries, but you only made Y’, rather than ‘your deliveries are slower than expected’.

Member States should also include explicit human review guarantees, rather than limiting review to when 'appropriate'. Where algorithms are used to make decisions in the workplace, there should always be a human either checking, and/or able to review, any decisions. As such, when adopting their own legislative frameworks we recommend that Member states require platforms to provide explanations in accessible language that both workers and their representatives can access, including updates on algorithmic or technical developments of the platform.

Finally, it is also essential that references to classification of employment not be interpreted to evade the obligation of Member States and platforms under this article. All workers in the platform economy, regardless of their employment status, should enjoy these rights.

More detailed provisions on transparency and explainability were contained in a non-binding Recommendation that was meant to be adopted alongside C193. However, this was dropped and not voted on at the ILC26.

Ratification and Recommendations

Adopting this historic Convention is a positive step forward for workers' rights, and the ILO should look to embed this legacy by securing wide Member State ratification. There is also a significant role to support the integration of the Convention at national level and ensure the Convention is effectively enforced, including by developing guidance and recommended practices to facilitate strong implementation with common elements across jurisdictions. The ILO, Member States and platforms alike should engage with workers, representatives, and NGOs to develop legislative frameworks in open and inclusive dialogue.

Importantly, the ILO must not lose sight of the importance of the Recommendation, which was not adopted earlier this year, and should bring it back for further discussion at a future International Labour Conference. The Recommendation would provide essential guidance to Member States and platforms not only to clarify and address some of the gaps identified above in relation to the language of the Convention, but also to support Member States and other stakeholders to understand and interpret some of the newer provisions of the Convention, which may pose new legislative challenges. Among other things, a new Recommendation can provide further details around the human rights implicated and risks experienced by workers, and it can further complement and bolster the Convention based on practical experience. The Recommendation could also provide a space for the ILO to consider how to take forward the groundbreaking provisions provided in the Convention around the need to regulate algorithmic management to other sectors and workplaces beyond the platform economy.

As such, PI recommends that:

  • States ratify Convention C193;
  • States consult with union representatives, civil society and independent technical experts on legislative measures to ensure effective national implementation and enforcement;
  • Digital platforms engage with workers' representatives and consult with civil society and experts to give full effect to the rights and obligations in the Convention;
  • The ILO provides technical expertise and guidance to support swift ratification and effective implementation and enforcement of the Convention.

Convention C193 has laid the groundwork that recognises the impact digital technology is having in the workplace. Our fight will continue for stronger safeguards for workers against the harms of automated decision-making in other spaces and forums.