Decent Work in the Gig Economy: An Appraisal of the EU and ILO Regulation of Digital Labour Platforms

Table of Contents: 1. Introduction: operation and scope of digital labour platforms. – 2. The risks for decent work in the gig economy. – 2.1. Fundamental principles and rights at work. – 2.2. Employment creation. – 2.3. Social protection and social dialogue. – 3. The EU Platform Work Directive. – 3.1. Legal qualification of people working via DLPs. – 3.2. The algorithmic management. – 4. The ILO project for a Convention on platform work. – 4.1. Is there a need for a specific regulation? – 4.2. The draft Convention. – 5. Conclusions.

Abstract: Digital Labour Platforms (DLPs) have drastically transformed the labour market over the past fifteen years, creating new and flexible job opportunities while also undermining the fundamentals of decent work. People who work via DLPs frequently experience unfair treatment, including discrimination, poor wages, and uncontrolled exploitation. This article explores the various responses to the phenomenon so far elaborated at the EU and international level, with the aim to emphasise their benefits and defects in terms of decent work. In particular, a thorough reading will be offered of both the recent 2024 EU Platform Work Directive and the upcoming ILO Convention on Decent Work in the Platform Economy, currently under debate before the ILO General Conference.

Keywords: Gig economy – digital labour platforms – decent work – social rights – EU Platform Work Directive – ILO Convention on Decent Work in the Platform Economy. 

1.   Introduction: operation and scope of digital labour platforms

The phenomenon of digitalisation consists of the ‘ongoing integration of digital technologies and digitised data across the economy and society’.[1] Thanks to the creation of digital platforms, multiple aspects of human life are gradually shifting from the material dimension to a virtual one. Social media platforms, like Facebook, allow people to create and cultivate human relations through digital interactions, whereas streaming platforms, such as Netflix, offer a wide variety of entertainment always available via the internet. With online shopping platforms, as Amazon, buying and selling can be easily done at a fingertip, and any products are delivered directly at home. Even education migrated to digital platforms with the establishment of telematic universities and online professional courses.

The digital advancement did not spare the labour field. The last fifteen years have seen a rapid and intense flowering of Digital Labour Platforms (DLPs), which has caused a radical transformation of the job market due to the convergence of certain social, economic and political factors.[2] To grasp the dimension of the phenomenon, it suffices to recall that the market value of the labour platform economy, known also as ‘gig economy’ or ‘sharing economy’, increased from 7.1 trillion US dollars in 2018 to 10.2 trillion in 2023.[3] In parallel, the number of labour platforms grew from 193 in 2010 to 1070 in 2023.[4] Focusing on the human capital, in 2022, the platform workforce amounted to 28 million people only in the territory of the European Union (EU), a number expected to almost double by the end of 2025.[5]

A DLP typically identifies a virtual space where either professional or non-professional services are offered, sold, and sometimes performed.[6] From a practical standpoint, these platforms aim at establishing a virtual connection between a group of users who are in need of a particular service, known as ‘requesters’ or ‘consumers’, and a group of ‘workers’ who offer paid services. Technically speaking, three distinct categories of DLPs are to be distinguished. On web-based platforms, the work assignments have a professional nature and are performed online or remotely by so-called ‘crowd workers’.[7] These platforms are a valuable tool for outsourcing micro or macro phases of articulated projects, varying from fact-checking, text creation, or audio translation, to more complex activities, as project design or software development.[8] Clickworker, Microworker, Crowdflower (now Figure Eight), or Amazon Mechanical Turk fall within this category. On the other side, the location-based – or app-based – platforms provide services, either professional or not, that are performed offline by ‘on-demand workers’.[9] This is the case of food delivery apps like Deliveroo, Glovo, or JustEat, as well as private transport services apps, such as Uber. The third model is that of content digital labour platforms, which are not based on the mere interplay between supply and demand, but instead provide the so-called ‘content creators’ with a virtual space where to share their intellectual products with the public.[10] The profit in this case is made through advertisement or subscriptions, and depends on the creator’s ability to build a following of audience. Unquestionably, the most famous example is that of YouTube. 

The rationale behind DLPs is to take advantage of the technological development for maximizing profits and opening the job market. From the businesses’ perspective, the assignment of highly specialised tasks to digital workers may reduce the cost for employees, infrastructures, transport, and communication without reducing productivity, as the platforms guarantee specific quantity and quality standards.[11] Replacing employees with digital workers may result in labour cost savings of up to 30%.[12] In parallel, outsourcing – or better hiding – tedious activities through web-based platforms allows big companies, as well as innovative start-ups, to build up their identity as technological and innovative entrepreneurs, which is definitely more captivating for attracting investments and funds.[13] Reversing the viewpoint, via DLPs, workers are granted the possibility to work at their convenience, meaning where, when, and how they prefer, particularly if web-based platforms are considered.[14] Such an extreme flexibility has a positive impact in terms of inclusivity of the job market, which becomes accessible also to people unable to leave their home or to afford a living in a different city, as people with disabilities, single parents responsible for childcare or young professionals.[15]

On a closer sight, however, the drawbacks of the gig-economy far outweigh the benefits. The essential requirements for decent work, as formulated by the International Labour Organisation (ILO), are constantly at risk of being jeopardised due to the functioning of DLPs, as will be further seen in paragraph 2. Setting aside the domestic responses to the phenomenon, which have been extensively and exhaustively covered in legal literature,[16] this article explores the regulation of DLPs so far elaborated at the EU and international level, with the aim to assess how and to what extent they contribute to restoring decent work in the gig economy. Paragraph 3 will dissect the main legal instrument so far developed by the EU in the field, namely the Platform Work Directive, whereas paragraph 4 will focus on the ILO’s legal initiatives in terms of DLPs, with particular regard to the project of the Convention on Decent Work in the Platform Economy, which is currently under debate before the ILO General Conference. Paragraph 5 will conclude by addressing the main differences between the EU and ILO approaches and their consequences in terms of final results. 

2.   The risks for decent work in the gig economy 

In 1999, in the throes of the economic globalization,[17] the ILO Director-General Juan Somavia first introduced to the 87th Session of the International Labour Conference the concept of decent work as one encompassing four strategic pillars of the ILO Agenda for the upcoming century: promoting the fundamental principles and rights at work; fostering employment; strengthening social protection; and improving social dialogue and tripartism.[18] A decade later, the four-pronged character of decent work was confirmed in the ILO Declaration on Social Justice for a Fair Globalization adopted by the International Labour Conference at its 97th Session, which was the third major statement of principles and policies since the ILO’s establishment.[19] In parallel, a Tripartite Meeting of Experts established by the ILO Governing Body pinpointed a list of indicators for measuring the four dimensions of decent work, thus contributing to outlining a uniform concept and offering a better understanding of its content.[20]These included, among others, the precarious employment rate, the working poverty rate, the excessive working time, and the share of people benefitting from a pension.[21]

From a broader perspective, the promotion of decent work, in combination with full employment and economic growth, has been listed among the 17 Sustainable Development Goals set by the United Nations (UN) 2030 Agenda, particularly as Goal 8. The multifaceted identity of the concept remains unchanged, as proved by the specific targets associated with the main goal.[22] Target 8.5 highlights the need to achieve full employment and equal pay for all, notwithstanding the gender, age, or health conditions of workers. Targets 8.7 and 8.8 address the fundamental principles and rights of workers, emphasising the safety of the working environment. Target 8.10 aims at improving social protection by expanding workers’ access to banking, insurance, and financial services. Goal 8 is therefore expected to be interpreted in line with the ILO’s main initiatives on decent work.[23]

Besides being mentioned in soft law texts, the concept of decent work has gained autonomous normative value as an individual right. The right to work codified at Article 6 of the International Covenant on Economic, Social and Cultural Rights (ICESCR) must indeed be interpreted as a right to decent work, according to the General Comment 18 of the Committee on Economic, Social and Cultural Rights of 2006.[24] The essential requirements for a work to be decent are detailed by the following norms, which the Committee defined as ‘interdependent’.[25] In particular, Article 7 confers general application to the right to fair wages, safe and healthy working conditions, equal opportunities, and limitation of working hours, whereas Articles 8 and 9 protect the freedom of association and the right to social security, respectively. 

 Nowadays, decent work is thought to live a ‘crisis-non crisis’ condition since its observance is broadly and firmly promoted, as the aforementioned texts demonstrate, but rarely followed by practical attempts of enforcement.[26] Focusing on the gig economy, which is the subject of this article, the following paragraphs will investigate the implications for decent work resulting from the operation of DLPs, taking into account the different pillars of the notion separately.

2.1.  Fundamental principles and rights at work 

The first pillar of decent work corresponds with the promotion of fundamental principles and rights a work, which are commonly known as core labour standards (CLSs). The 1998 ILO Declaration on Fundamental Principles and Rights at Work launched the concept and encompassed within the category of CLSs four principles: the freedom of association and collective bargaining, the abolition of forced labour, the abolition of child labour, and the prohibition of any form of discrimination at work.[27] Moreover, in 2022, the 110th Session of the International Labour Conference opened the catalogue to include also the right to a safe and healthy working environment.[28] All the 187 member States of the ILO are expected to observe the CLSs, irrespective of whether they ratified the ILO Conventions on the same subject matter.[29]

Having regard to the phenomenon of DLPs, the CLSs mostly at risk of being jeopardised seem to be the freedom of association, the prohibition of discrimination, and the right to a safe and healthy working environment. Whereas the risks related to the former are contingent upon the legal qualification of platform workers as established by the Terms of Service (ToS) agreements of DLPs – an aspect which will be further investigated below when discussing the social dialogue pillar – the discriminatory practices as well as the hazardous or unhealthy working conditions are a consequence of the practical operation of DLPs, which essentially consists of algorithms matching requesters and workers basing on a rating system. 

The opaque, and sometimes unreliable, functioning of the algorithms threatens fair competition between digital workers.[30]Algorithms’ design features and source code are hardly disclosed to users and almost impossible to be inferred from practice.[31]Within this framework, both direct and indirect forms of discrimination are not unusual. For instance, assigning a higher rating to workers on the basis of their availability to work on festive days may lead to indirect religious discrimination. The same holds true for workers’ availability to work at night or in the evening, which impairs people in charge of childcare. As for direct discrimination, studies on micro-task crowd work highlight that people from developing countries are systematically precluded from applying for the most complex and, thus, well-paid, job offers.[32] Similarly, providing users with different treatments depending on the subscription to a ‘premium package’ leads to discrimination for many workers who are unable to bear such a cost.[33]

Alongside having discriminatory effects, the rating system applied by the algorithms frequently builds a conditional relationship between digital workers and the platforms. To secure and safeguard a high-ranking position, workers are indeed expected to be always connected, and most importantly, always available to accept new tasks. Therefore, it is common for them to work on weekends, late at night, in adverse weather, or even in poor health conditions. Disconnecting from the platform when demand peaks, on the contrary, is likely to result in a forcible and unmotivated deactivation of the account,[34] thus leaving digital workers without any revenue as well as any social protection. Given these circumstances, the working conditions in the gig economy appear to be unhealthy, from both a physical and a psychological perspective.[35] Digital workers accept and undergo a stressful working life simply because this is the inescapable condition to stay in the digital market. That being said, DLPs’ promise of flexibility proves to be deceptive.

2.2.  Employment creation

Decent work’s second pillar, namely that of employment creation, has a twofold meaning. It is firstly connected to the enlargement of the job market by either creating new job opportunities or opening the existing ones to unemployed people.[36] In this respect, platform work apparently meets the expectations. Thanks to DLPs, people can now access a countless catalogue of job offers from all over the world and carry out paid assignments easily from home with a simple click on the laptop or smartphone. At the same time, small gigs that were formerly confined to teenagers or part-time workers, like food delivery, became real job opportunities open to a wide range of people, as no particular training or organisational structure is needed. 

On the other side of the coin, however, working in the gig economy is by definition precarious. DLPs encourage over-recruitment as a market strategy, thus making the digital workforce a fungible resource where workers’ competition becomes vital, and the only way to avoid replacement is to fully dedicate – or better subordinate – to the platform, as seen above.[37] In light of these considerations, several voices claim that the gig economy echoes the typical features of the industrial age, rather than bringing any true innovation in the labour field.[38]

Secondly, the creation of employment must always be accompanied by the attribution of a fair salary to all workers, regardless of whether they qualify as employees or as independent contractors.[39] Such an aspect turns out to be particularly critical in the gig economy, where digital workers usually settle for low wages, which may even fall below the poverty threshold when vulnerable categories of workers are taken into account, like women or migrants.[40] According to the ILO, the average hourly wage on web-based platforms amounts to 3.4 US dollars, with 66% of workers earning less[41] and 37% of workers relying on crowd work as their primary source of income, with critical results in terms of capacity to afford social protection and health coverage.[42] In addition, on certain DLPs, the payments are not monetary, but consist of gift vouchers, a circumstance that threatens the workers’ capacity to meet the daily needs, particularly in developing countries.[43] The scenario does not change when on-demand workers are considered. The data collected by ILO, in fact, displays a complex situation with hourly earnings fluctuating between 1 to 3-4 US dollars, save for very sporadic exceptions of 7-8 US dollars in a few countries, such as Chile and Lebanon.[44]

The poverty of wages depends on multiple factors. First, the commission fees on DLPs, besides being significantly high, are mostly unloaded on workers.[45] Second, the time that digital workers spend in invisible labour, such as managing the payments, earning additional qualifications to be competitive on the market, or searching for new tasks, is unpaid and amounts to approximately twenty minutes per hour.[46] Third, digital workers – particularly, on-demand workers – are often responsible for work-related costs. For instance, they incur daily transport expenses, such as those related to vehicles and fuel, without any chance to get a refund from the platforms. Finally, on certain platforms such as Amazon Mechanical Turk, unsatisfied consumers are recognised the right to reject the product obtained through the platform, without the need to provide any motivation and, more importantly, without the obligation to fulfil the payment. This opens to the risk of legitimising a ‘wage-theft’,[47] against which workers are completely unarmed since the ToS agreements regularly assign the intellectual property rights over the final products to the requesters.[48]

2.3.  Social protection and social dialogue

The third pillar of decent work corresponds with the social protection of workers, which in turn includes pension payment, unemployment and invalidity insurance, and maternity and sick leave, at least.[49] The social dialogue, instead, which is the fourth requirement for a work to be deemed decent, depends on workers’ capacity to exercise their freedom of association by joining trade unions or informal associations, and is commonly expressed through collective bargaining, one of the core labour standards mentioned above. 

In the gig economy, the major obstacle to the enforcement of both these aspects lies in the legal qualification of digital workers. The ToS agreements that DLPs unilaterally arrange and that users are requested to subscribe in order to register an account usually qualify digital workers as ‘independent contractors’ or ‘self-employed workers’, with some exceptional cases where the legal terminology is even avoided in favour of informal labels, like ‘participants’, ‘partners’, ‘drivers’, or ‘captains’. As a result, similarly to other forms of non-standard employment,[50] platform workers are systematically denied the rights that domestic legislations subordinate to the existence of an employment contract. This is true in most States for both the social security coverage and the freedom of association indispensable for collective bargaining.[51] Hence, social protection remains an individual responsibility of workers, who, however, are quite unlikely to autonomously afford this cost, given the low wages they are used to.[52]

The impairment to the third and fourth pillars of decent work appears to be much more severe if one considers that the legal qualification provided by the ToS agreements is frequently fictitious, and digital workers are de facto treated as employees. More precisely, DLPs always exert indirect control upon digital workers through reputation systems: consumer reviews and feedbacks are among the criteria processed by the algorithms to determine workers’ rating and, consequently, they have an impact on the opportunities for future assignments.[53] In addition, digital workers can undergo direct control by the platforms, particularly when they receive precise instructions to be observed in delivering the tasks, are provided with working tools, or are monitored by the platforms while performing the services, all these aspects being signs of organisational integration.[54] By way of example, the platform Lyft, a competitor of Uber in the field of transport services, provides its drivers with a set of guidelines to be observed, which include how vehicles must be cleaned, how drivers must welcome the passengers, and who can be hosted on board.[55] Web-based platforms, moreover, are often equipped with a timer to monitor the working hours and, in certain cases, they even take regular screenshots of the workers’ display in order to verify whether they are fully dedicated to the task assigned.[56]

Furthermore, even though a proper exclusivity clause rarely appears in the ToS agreements, DLPs incentivise long-term work commitments by offering benefits in case of prolonged activities and by denying, or at least hampering, the workers’ ability to ‘migrate’ their rating on a different platform ‘lock-in effect’.[57] Within this scenario, the principles of freedom and flexibility, which should inspire the functioning of DLPs, are ineffective, and the digital worker ends up being an employee in disguise of the platform, which in turn becomes a service provider to the requester.[58] Even when the control element is difficult to prove, platform workers have been demonstrated to suffer from the vulnerabilities common to employees.[59]

In order to face these issues, domestic jurisdictions opt for the requalification of platform workers as employees whenever the essential criteria for such a legal status are de facto satisfied, consistently with the ‘primacy of facts’ principle prescribed by the ILO Recommendation 198. With a decision of 2018, the California Supreme Court recognised in favour of Dynamex drivers a presumption of subordination for wage orders, shifting upon the platform the burden to prove the contrary.[60] Afterwards, platform workers have been directly re-qualified as employees by the French Cour de Cassation in the TakeEatEasy case of 2020[61] and, in the same year, by the Spanish Tribunal Supremo in the Glovo case,[62] in both circumstances with the argument that the practical control of DLPs on workers and their exclusion from management and profits revealed a subordinate relation. The same epilogue has been registered in further European States, such as Ireland,[63] the Netherlands,[64] Germany,[65] and, in some cases, Italy.[66] This solution, however, gives uncertain results insofar as it depends on the judicial sensitivity and the factual backgrounds of each case. In this regard, it suffices to recall the British jurisprudence, where Uber drivers have been qualified as ‘workers’ whereas Deliveroo riders have not.[67]

More incisively, some voices in legal literature have suggested embracing a human rights approach and amending domestic legislation accordingly, particularly by extending to all workers the same legal treatment as employees.[68] A normative reform of this kind is actually in line with the international conventions on social rights of both universal and regional character, where social protection and freedom of association are rights due to everyone.[69] In this respect, it bears noting that in 2021 the European Committee of Social Rights posed a targeted question to 33 member States on the observance of Article 12, paragraph 3, of the European Social Charter, asking in particular to report on the social security coverage for platform workers. Interestingly, the Committee has not been able to adopt a final conclusion on the matter as the majority of States provided no useful information, a circumstance which ‘suggests that States are lagging behind in developing their social security systems in response to changes in the labour market’, as the Committee itself concluded.[70] In parallel, with the Advisory Opinion OC-27/21, the Inter-American Court of Human Rights emphasised States’ obligation to guarantee the freedom of association and the right to collective bargaining, and warned that ‘the protection of these rights must be understood in consideration of the fact that labour relations are in constant flux due to a variety of factors, including the use of new digital technologies in the workplace’.[71] As a result, ‘States must adopt legislative and other types of measures focused on individuals, and not primarily or exclusively on markets, that respond to the challenges and opportunities brought about by digital transformation of work, including work over digital platforms’.[72]

3.   The EU Platform Work Directive 

Among the multiple EU acts addressing, either directly or indirectly, the consequences stemming from technological development in terms of social, particularly labour, matters,[73] a concrete attempt to meet the new challenges of the gig economy has been made by the Directive (EU) 2024/2831 (Platform Work Directive, or Directive),[74] which was published in the Official Journal on 11 November 2024 and eventually came into effect on 1 December 2024, following a long negotiation process.[75] The purpose of the Platform Work Directive is to enhance the working conditions of platform workers through the adoption of common parameters for their legal qualification, while in parallel ensuring a transparent functioning of the algorithms and the safe treatment of users’ personal data.[76] The following paragraphs will investigate both these aspects separately, with particular attention to the consequences in terms of decent work. 

Notwithstanding the broad ambitions of the Directive, its scope of application undergoes some restrictions ratione personae. A category of platform workers, namely that of content creators, is excluded from the scope of application as the definition of ‘digital labour platform’ set by Article 2 is not matched in this case. In particular, the requirement of organisation of work by the platform is missing.[77] From a ratione loci perspective, instead, a broad scope of application is set by Article 1, paragraph 3, according to which the Directive applies whenever the platform work is performed in the territory of the EU, irrespective of the place of establishment of the DLPs involved or of any other law otherwise applicable.

3.1.  Legal qualification of people working via DLPs 

Under Article 2, the Directive makes a distinction between a ‘person performing platform work’ and a ‘platform worker’, with the latter being the one who ‘has or is deemed to have an employment contract or an employment relationship as defined by the law, collective agreements or practice in force in the Member States with consideration to the case law of the Court of Justice’.

In consideration of the practical consequences stemming from the legal qualification, Article 4, paragraph 1, compels Member States to ‘have appropriate and effective procedures in place to verify and ensure the determination of the correct employment status’, then referring at paragraph 2 to the primacy of facts principle as the rule guiding such procedures. In order to overcome the risk for incoherent outcomes that is inherent to States’ autonomy, Article 5, paragraph 1, introduces a legal presumption in favour of the employment relationship ‘where facts indicating direction and control’ are found, saving the possibility for DLPs to rebut it. These aspects must be valued in light of domestic laws, collective agreements, national practice, and eventually the jurisprudence of the Court of Justice of the EU (CJEU). Moreover, according to paragraphs 2 and 3, Member States shall codify and apply the legal presumption in administrative and judicial proceedings, whereas its applicability to tax, criminal or social security matters is left to States’ discretion. Extending the legal presumption to the social security field would have provided a precious contribution to the third pillar of decent work, which instead remains generally underexplored by the Directive, as will be further seen below. Such a legislative choice is puzzling, mostly because Article 153 of the Treaty on the Functioning of the EU (TFEU), which is one of the Directive’s legal bases, includes social security among the matters of shared competence between the EU and Member States.[78]

Originally, in the Directive’s first draft as elaborated by the Commission, the legal presumption was equipped with five criteria for examining the control link between digital workers and platforms, with the requirement of a positive conclusion when two of them were present.[79] These criteria sharply emphasised some of the essential features of platform work, even though they were not adequate to reflect the characteristics of all different types of DLPs.[80] A first sign of control was the platforms’ determination of workers’ wages, as well as the performance regulation and supervision. In addition, any direct or indirect restriction of workers’ freedom to organise their own work was deemed relevant, in particular workers’ incapacity to determine their time schedule, to refuse tasks, and to delegate the performance to third subjects. A final criterion was the subsistence of any exclusivity in the relation between platforms and their workforce.[81] In the wake of the criticism made by the European Economic and Social Committee,[82]the European Parliament’s Special Rapporteur Gualmini proposed the suppression of the entire provision in May 2022, a position which was eventually agreed and merged in the Directive’s text despite an initial opposition of the Council.[83]

On the one hand, establishing uniform criteria for interpreting the control link and, thus, applying the employment presumption would have undoubtedly contributed to outlining a European definition of ‘employment contract’,[84] that differs from the one the CJEU has historically used in the common market[85] and, instead, draws from the practices of domestic labour systems, with promising harmonisation prospects.[86] On the other hand, however, the criteria set by the Directive were inconsistent with certain national laws,[87] could have been easily circumvented by DLPs through a business reorganisation,[88] and, above all, were viewed by a number of Member States – with France on top – as an unacceptable interference into domestic competence. In a nutshell, the practical interest of making the Directive a reality overshadowed the original goal of creating a discipline as uniform as possible. 

In light of these premises (and promises), the final word of the Directive appears to be a hollow victory. During the re-qualification procedure, in fact, platform workers are still expected to prove that both direction and control exist in order for the presumption of employee status to be applied.[89] Even though the benchmarks are not restricted to domestic laws, but extend to the jurisprudence of the CJEU, practical experience shows how challenging it may be to provide solid evidence of the direction and control factors, which more and more frequently are disguised in the algorithm’s operation.[90] In addition, in the case of cross-border relations, some concerns arise with regard to the law applicable to requalification procedures. First, the absence of a harmonised definition of employee leaves it uncertain whether to apply Article 8 of the Regulation (EC) No 593/2008 (Rome I),[91]which refers exclusively to ‘individual employment contracts’. Second, even admitting that Article 8 applies, the risk for social dumping stemming from the choice of law made in the ToS agreements is neutralised according to paragraph 1 of Article 8 only within the EU borders. Conversely, the applicability of the labour law of third States is a priori excluded by Article 1, paragraph 3, of the Directive, even when it provides platform users with higher protection.[92]

3.2.  The algorithmic management 

In the Platform Work Directive, Chapter III dedicates a major space to the discipline of algorithms, which are distinguished in two different categories depending on whether they aim at supervising users’ performances (monitoring systems) or at elaborating decisions having regard to the users’ treatment (automated decision-making) systems. In this respect, it bears noting, the Directive embraces a techno-determined approach.[93] The opportunity to include the algorithmic technology within the functioning of DLPs, in fact, is not discussed nor opened to negotiation with social parties’ representatives. Rather, the Directive assumes that algorithms are essential and limits to regulate their operation, having regard to three different aspects: transparency, data protection, and individual rights. Notably, the two types of algorithms commonly undergo the same discipline, which does not come as a surprise considering that they typically operate together. A further distinctive feature of Chapter III relates to the addressees of protection, which mostly correspond with persons performing platform work. By so doing, employees and self-employed are de facto granted the same treatment, a choice that legal literature has praised as a novelty and an important achievement in the path of digital workforce protection,[94] particularly if compared with the previous Directive 2019/1152 on transparent and predictable working conditions.[95] On a closer look, however, the few prescriptions of the Chapter having a real impact on substantive labour rights are limited to platform workers and, more importantly, do not deal with some essential aspects of decent work. 

Having regard to transparency, Article 9, paragraph 1, of the Directive requires DLPs to inform persons performing platform work, as well as their representatives and, upon request, public authorities, about the existence of algorithms of both types, the data involved in their operation, the criteria followed in processing them, and eventually the people entrusted with access to those data, a prescription which is a step ahead of the Regulation (EU) 2024/1689 on artificial intelligence (AI Act).[96] Further guarantees are ensured when platforms’ decisions go to the detriment of people performing platform work. It is the case of decisions to restrict, suspend, or terminate the account, as well as decisions to refuse the payment or to modify the contractual status of users. Under Article 9, paragraph 1, letter b), number iv) of the Directive, the grounds for pronouncements of this kind must be disclosed in advance. Additionally, Article 10, paragraph 5, establishes that a final intervention by ‘a human being’ is always necessary.[97] In the same vein, any detrimental decision must be equipped with a written statement of the reasons underpinning it under Article 11 of the Directive. Insofar as they discourage the practice of unjustified termination of platforms’ accounts, these provisions represent a useful, albeit indirect, contribution to the second pillar of decent work, particularly from the perspective of job opportunities. 

A second crucial aspect of Chapter III relates to the use and treatment of individuals’ personal data for the purposes of platforms’ functioning, which operates as lex specialis in relation to the Regulation (EU) 2016/679 on data protection (GDPR).[98] Article 7 of the Directive bans DLPs from collecting ‘any personal data of a person performing platform work while that person is not offering or performing platform work’ (letter c). In parallel, algorithms cannot process any personal data related to the emotional and psychological state of the individuals (letter a), as well as any data extracted from private conversations (letter b). Consistently with Article 9 of the GDPR, the remaining letters of Article 7, which were absent in the first draft, restrict the use of personal data for improper purposes, namely to predict the exercise of rights (letter d), to infer sensitive information such as racial origin, political opinions, religious beliefs, health status or sexual orientation (letter e), and eventually to determine the biological identity of individuals (letter f). By limiting access to sensitive data, the Directive safeguards people performing platform work against direct discrimination and, therefore, enforces one of the core labour standards.[99] Conversely, no guarantees are offered against indirect discrimination, which unquestionably represents the most frequent case and includes distinctions based on workers’ time availability and geographic location, as seen above in this article. As for the treatment of data, DLPs are always required to perform a data-protection impact assessment with persons performing platform work and their representatives to be involved in the procedure, an aspect that makes such a prescription much more incisive than the GDPR itself.[100] Finally, Article 9, paragraph 6 recognises to persons performing platform work ‘the right to the portability of personal data generated through their performance of work’, a right whose exercise DLPs are expected to facilitate by providing adequate tools.

A last profile to be considered is that of labour rights. Article 10, paragraph 1, of the Directive codifies an obligation of regular human oversight upon the impact of algorithms on persons performing platform work, their working conditions, and the equal treatment at work. If any risk of discrimination or any breach of rights is detected, DLPs are expected to take the necessary steps to correct the defects under paragraph 3, including modifying and discontinuing the algorithm. In this respect, however, a strong asymmetry surfaces between employees of DLPs and people performing platform work in general. Only the representatives of the former are involved in the oversight procedure, whereas, according to Article 15, the representatives of the latter are only admitted to receive information on the evaluation once concluded. The same asymmetry extends to substantial labour rights. In particular, Article 12, paragraph 1, requires DLPs to evaluate the consequences that the use of algorithms may determine on the health and safety of platform workers, and to introduce appropriate protective measures to face these risks, whereas paragraph 3 prohibits DPLs from using algorithms ‘in a manner that puts undue pressure on platform workers or otherwise puts at risk the safety and physical and mental health of platform workers’. Two core labour standards are therefore openly addressed by the Directive, namely the prohibition of discrimination and the right to safe and healthy working conditions. Yet, as long as their implementation is dependent on the existence of an employment contract, there does not appear to be any discernible improvement in terms of advancing decent work in the gig economy.

An even more meagre result is reached by the Directive in terms of collective labour rights, which are mostly disregarded or subordinated to the existence of an employment relation, as seen above. According to Article 13, DLPs must inform and consult only with platform workers’ representatives, even in the case of technical decisions, such as the introduction or modification of platforms’ algorithms.

Aside from this prescription, social parties bear no role in the discipline of platform work set by the Directive, a choice that has met with large criticism in legal literature.[101]

That being said, multiple essential requirements for a work to be considered decent found no room in the Platform Work Directive. Save for a brief reference to the European Pillar on Social Rights in the Preamble, the right to social security coverage is entirely disregarded, remaining a consequence of the legal qualification of people performing platform work, which in turn largely depends on Member States’ discretion as seen in the previous paragraph. The same goes also for the freedom of association and the right to collective bargaining, as well as for the right to minimum wage, all of them being absent from the Directive’s text. The legislative choice not to address the last two profiles is formally justified under Article 153, paragraph 5, TFEU, as they fall outside EU competences. That notwithstanding, it has been duly emphasised in legal literature that, in different circumstances, the EU addressed these matters by collateral instruments,[102] an example being the Minimum Wages Directive, which the EU adopted in 2022 on the basis of its competence to regulate working conditions under Article 153, paragraph 1, letter b), TFEU.[103] As for the freedom of association, the exclusion made by paragraph 5 of Article 153 must be balanced with the provision of paragraph 1, letter f), of the same Article, which includes collective representation and defence of labour rights within EU competence.[104] That being said, the EU has apparently been unwilling to address the most delicate issues of decent work in the gig economy, rather than being unable to do so. 

4.   The ILO project for a Convention on platform work 

Moving to the international level, the ILO seems to be fully aware of the main obstacles to decent work that are caused by the emergence of DLPs. After having carried out two surveys among thousands of crowd workers and on-demand workers in 2015 and 2017, the 2019 Centenary Declaration on the Future of Work invoked the adoption of policies which ‘respond to challenges and opportunities in the world of work relating to the digital transformation of work, including platform work’.[105] The findings of the previous surveys have then been discussed in light of the international discipline of labour rights in the 2021 Report World Employment and Social Outlook, the first one specifically devoted by the ILO to the topic of platform work.[106]

As a result, during the 347th Session of the ILO Governing Body, held in Geneva between 13 and 23 March 2023, it was decided to place in the Agenda of the 113th Session of the General Conference an item on decent work in the platform economy.[107] The purpose is to adopt a standard-setting instrument on the basis of a double discussion, with a first general round of debate held in 2025 and a second final one expected in 2026. 

The following paragraphs will investigate why such a project appears necessary despite the broad normative framework already in force within the ILO and whether it will effectively contribute to enhancing decent work for platform workers. 

4.1.  Is there a need for a specific regulation? 

As clarified in the first Report developed for the 113th Conference’s Session and titled Realizing Decent Work in the Platform Economy, the introduction of a special normative instrument on the subject of platform work is necessary, being the existing legal framework inadequate to address the main concerns in the field.[108]

On the one hand, it is undisputed that the five core labour standards elaborated by the ILO, which safeguard the main labour rights and are currently recognised as being customary, apply to an all-embracing category of workers, including both employees and self-employed people. The same goes for the corresponding ILO Conventions dealing with fundamental labour rights.[109]Consistently, in 2020, the ILO Committee of Experts on the Application of Conventions and Recommendations concluded that ‘the full range of fundamental principles and rights at work are applicable to platforms workers in the same way as to all other workers, irrespective of their employment status’.[110]

On the other hand, however, some personal and material gaps in the current legal framework have been detected by the 2024 Report. Alongside the lack of several essential rights – such as minimum wage, social security, and protection from the unlawful termination of contract – which ILO Conventions traditionally grant to employees only,[111] there are many features of platform work that escape the ordinary discipline of labour rights so far elaborated within the ILO. These primarily relate to the technological character of DLPs, especially the use of algorithms to allocate tasks and supervise their fulfilment, and the protection and portability of the personal data collected when platform workers are online.[112] Another peculiar aspect of platform work concerns the time spent in invisible labour, namely the time when crowd-workers and on-demand workers are logged in to DLPs searching for new jobs or waiting for new assignments, but without getting any remuneration. Such a ‘waiting time’ can hardly be qualified as ‘hours of work’ pursuant to the Hours of Work Convention of 1930 (ILO Convention No. 30), and even if so, this would apply only to employees.[113]

Finally, working through web-based platforms typically assumes cross-border proportions, considering that an efficient internet connection is all that is needed to obtain and perform the most different tasks, with the result that a job can possibly be outsourced all over the world. 

4.2.  The draft Convention

In the light of the survey conducted among States and social parties in the second half of 2024 and summarised in the second Report Realizing Decent Work in the Platform Economy of 2025, the ILO Governing Body concluded in favour of introducing new standards on platform work in the form of a Convention supplemented by a Recommendation, and elaborated a preliminary set of conclusions in this respect.[114] After a first round of discussion between 2 and 13 June 2025, on the occasion of the 113th Session of the General Conference, the form, the definition, and the scope of the standards have been approved.[115] A first draft of the Convention was then released on 15 August 2025,[116] postponing the determination of the final content to a second round of discussion, to be held during the 114th session of the General Conference in 2026.

Despite not being definitive, it is worthwhile making some considerations on the Convention’s draft, in order to emphasise its innovative features in terms of decent work. The following pages will refer to the draft articles of the Convention and, when relevant, to the draft Recommendation annexed.

Starting from the crucial issue of the relation between legal qualification and substantial protection, the ILO seems oriented towards an original approach essentially consisting of separating the two aspects.[117] The traditional dichotomy between self-employed and employee is preserved, with States required to adopt adequate measures to ensure the correct classification of people engaged in platform work based on the primacy of facts principle (draft Article 10). At the same time, however, the dichotomy is overcome for the purpose of the protection of the main substantial labour rights, thus making a significant step forward compared to the Platform Work Directive. The Convention addresses ‘digital platform workers’ in general (draft Article 2, letter b), defined as any person engaged to work for the provision of service organised or facilitated by DLPs and for remuneration, ‘regardless of their classification of status in employment’ (draft Article 1, letter b). Therefore, the risk of gaps in the protection or unequal treatment that comes with subordinating labour rights to the legal qualification, as seen above, is definitively avoided, without jeopardising the status of those platform workers who legitimately claim self-employed status and the treatment that goes along with it at the domestic level. Against this framework, no attempt is made to elaborate a harmonised notion of employment for the purpose of platform work. This comes as no surprise considering the difficulties met by the EU Platform Work Directive in this respect and, more importantly, the uselessness of a common definition of employment insofar as the application of labour rights to people working via DLPs does not depend upon it.

Delving into the details of the substantial protection ensured to platform workers, the ILO’s project much outweighs the tepid results of the Platform Work Directive. Draft Article 3 compels States to take ‘measures to ensure that digital platform workers enjoy the fundamental principles and rights at work’, thus strengthening the general obligation to safeguard the core labour standards and giving substance to the first pillar of decent work. Two standards are addressed in detail, namely the freedom of association and the right to a safe and healthy work environment.

According to draft Article 3, paragraph 2, States shall protect the right of digital workers to establish or join organizations of their choosing. In this respect, the ILO Convention provides an effective contribution to creating and enhancing the social dialogue in the gig economy, an aspect which, on the contrary, was completely disregarded in the Platform Work Directive. Further practical indications for implementing the freedom of association are provided in the draft of the Recommendation that will accompany the Convention. States are indeed encouraged to create an enabling environment for social dialogue (paragraph 2 of the draft Recommendation), to strengthen the capacity of representatives to effectively defend the interests of the associations’ members (paragraph 3 of the draft Recommendation), and to ensure a transparent exchange of information for negotiating purposes (paragraph 5 of the draft Recommendation). Unquestionably, within the borders of the EU, these prescriptions could encounter some obstacles, considering the uncertain legitimacy of self-employed collective agreements under EU competition law.[118]

Having regard to the occupational safety and health, moreover, under draft Article 4 of the Convention, States shall require DLPs to make risk assessments, as well as adopt preventive measures and take control measures in order to avoid ‘occupational accidents, occupational diseases and any other injuries to health arising out of, linked with, or occurring in the course of work’. Once again, the ILO is taking a step further than the EU. Besides covering all people engaged in platform work – and not only the employees as seen in Article 12 of the Platform Work Directive – the ILO draft Convention codifies also a precise obligation for DLPs to provide workers with an adequate training in occupational safety and health as well as with an adequate equipment (draft Article 5), and recognises platform workers the right to abandon any work situation deemed to be dangerous (draft Article 6). Additionally, draft Article 8 introduces specific safeguards against violence and harassment in favour of digital platform workers.

Going beyond the catalogue of core labour standards, the upcoming ILO Convention openly extends to all platform workers the rights to remuneration, social security, and protection from unlawful termination of contract, thus filling the personal gaps that characterise the existing international normative framework, as seen in the previous paragraph. According to draft Article 11, States should ensure that the remuneration due to digital workers is adequate, paid in legal tender – or in kind, should domestic laws authorise it –, and paid in full and on time. The draft Recommendation completes the discipline by ensuring to digital workers a remuneration ‘at least equivalent to the statutory or negotiated minimum wage that is applicable to other workers in comparable situations’ (paragraph 10 of the draft Recommendation), prohibiting DLPs from charging fees or costs to digital workers (paragraph 13 of the draft Recommendation), and requiring States compensate the ‘waiting time’ (paragraph 14 of the draft Recommendation). This major challenge of platform work appears thus to be efficiently addressed, as opposed to the EU Directive, which disregards it. 

Moreover, a general obligation to provide digital platform workers with social security protection no less favourable than people working in similar circumstances is codified under draft Article 14. As for the termination of contract, draft Article 20 prohibits the suspension or deactivation of accounts on DLPs ‘when it is based on discriminatory or otherwise unjustified grounds’. Conversely, digital platform workers are to be recognised the right to decline tasks or disconnect without suffering any retaliation under paragraph 16 of the draft Recommendation, a prescription that echoes Article 22 of the Platform Work Directive. Thanks to these provisions, both social security and employment creation pillars are efficiently addressed to the benefit of all platform workers, whose legal treatment becomes consistent with the essential requirements of decent work.

Taking into account the technical features characterising DLPs, the ILO Convention, like the EU legislator before, makes reference to both the use of algorithms and the protection of workers’ data. The former is regulated by an obligation to transparency. According to Draft Article 15, in particular, DLPs must inform platform workers about the use of automated systems as well as about the impact they have on the working conditions and the access to work. In this respect, the draft Recommendation, at paragraph 22, encourages DLPs to regularly carry out an impact assessment. The ILO Convention remains at a general level, not providing a detailed list of the information to be disclosed on the functioning of the algorithms, as is instead done by the EU Directive at Article 9. However, a more innovative stance than the Directive is embraced on the matter of discrimination, with draft Article 16 expressly establishing that the use of automated systems by DLPs shall not infringe fundamental labour rights, and, in particular, shall not ‘result in any direct or indirect discrimination against digital platform workers’. Draft Article 18, then, generally prescribes the protection of personal data by recalling the observance of international instruments on the matter, whereas draft Article 19 provides a list of data whose collection, storage, processing, and use is completely forbidden. Such a list perfectly matches that made by Article 7 of the Directive. Additionally, under paragraph 26 of the ILO Recommendation, States are encouraged to introduce specific policies for the portability of personal data.

A final aspect to be considered is that of remedies. Draft Article 17 mandates that any automated system-based decision that affects working conditions or access to work must be accompanied by an explanation and, if it involves the refusal of payment or the suspension or deactivation of the account, be subject to human review. Draft Article 24 adds that platform workers are to be granted access to ‘safe, fair and effective dispute resolution mechanism and to appropriate and effective remedies’. The draft Recommendation eventually tackles the issue of trans-boundary relations, establishing that platform workers should be recognised the right to seek justice either in the State of their residence or in the State where the task is performed, independently from the place of establishment of the DLP involved in the proceedings (paragraph 29 of the draft Recommendation). At the moment, the possibility of claiming corporate responsibility for the breach of labour rights before an international forum is overlooked. Nonetheless, nothing excludes that a step in this direction could be made in the second round of discussion in 2026. If so, the upcoming ILO Convention would enhance not only the substance but also the efficacy of the protection offered to platform workers.

5.   Conclusions

The use of DLPs has drastically transformed the labour market over the past fifteen years, creating new and flexible job opportunities while also undermining the fundamentals of the idea of decent work. People who work via DLPs frequently experience unfair treatment, including discrimination, poor wages, and uncontrolled exploitation. In parallel, the legal qualification as self-employed workers unilaterally established by the ToS agreements of DLPs typically prevents platform workers from benefitting from social security coverage and contributing to social dialogue by joining trade unions and participating in collective bargaining. 

The solutions offered by the diverse legal orders to face such a phenomenon vary in nature and effectiveness. Focusing on the EU and ILO recent initiatives on the matter, this article has explored whether and to what extent they are adequate to ensure decent work for people involved in the gig economy. 

With the Platform Work Directive of 2024, the EU restates the interconnection between the legal qualification of platform workers and the recognition of substantial labour rights. At the same time, an attempt is made to harmonise domestic legislation through the codification of a legal presumption of employment status, which, however, turns out to be unsuccessful given the absence of uniform criteria for applying it. The only rights the Directive indiscriminately extends to all persons performing platform work are those connected to the algorithmic management, especially the right to be informed of the existence and functioning of the algorithms and that to the protection of personal data. By so doing, a tepid contribution to the enforcement of core labour standards is indirectly granted, particularly from the perspective of the prohibition of discrimination and the creation of safe and healthy working conditions. Conversely, the major concerns for platform workers are entirely disregarded. No mention at all is indeed made of social security and social dialogue, which remain circumscribed to people recognised as employees of the platforms. The same goes for the minimum wage, which is a fundamental requirement for decent work. Overall, the impression is that of a tentative progress in the social rights path or, to put it another way, of a missed opportunity to advocate for a genuine change in the regulation of the platform economy. 

A very different approach has been embraced by the ILO in the Convention on Decent Work in the Platform Economy, currently under discussion before the General Conference and expected to be finalised by 2026. A preliminary analysis of the draft reveals an intention to preserve the classical dichotomy between employees and self-employed under domestic law, but to abandon it for the purposes of platform workers’ protection. A broad list of substantial labour rights is thus recognised for all people working via DLPs, independently of their legal status. These include either core labour rights, such as freedom of association and occupational safety and health, or conventional rights, like social security, right to remuneration, and protection from the unlawful termination of contract. Besides filling some crucial gaps in the ILO’s legal framework, such an approach has the potential to offer a tangible solution to the main challenges connected with the spread of the platform economy, thus restoring the concept of decent work in this particular but controversial field. It remains to be seen whether States will be ready to follow the ILO down this road.