Trade union toolkit
A 2025 report from the OECD reveals that just under 80% of workplaces in Europe are using at least one algorithmic management system. At the same time, reports of operational failures, injustices to workers’ rights, health and safety pressures, and workplace culture erosion from algorithmic management are on the rise. Trade unions must shape and challenge if and how algorithmic management systems are developed and deployed in workplaces. To do so, trade unions can combine their labour expertise with collective bargaining, data access, and lobbying efforts to ensure algorithmic management systems improve productivity and efficiency (where beneficial and safe), while upholding effective worker protection.
In our toolkit for trade unions, we provide a summary of some of the existing levers in current laws and regulations that can be drawn upon by organisations to uphold, protect, and enhance workers’ rights. Here, we focus on two elements at trade unions’ disposal: information and consultation rights alongside data access rights. Following the enactment of the EU Platform Work Directive, platform (or ‘gig economy’) workers, long unprotected by the law, now find themselves with some of the strongest protections against intrusive and extractive algorithmic management practices.
As shown in our toolkit, protections do exist for ‘traditional’ workers, though they are weaker. The EU's Platform Work Directive can provide trade unions with a starting point on what to advocate for with respect to improved protections for workers against algorithmic management. This toolkit focuses on laws as they exist now. For trade union representatives looking for gaps in current protections and normative solutions, explore our policymaker toolkit.
Our toolkit summarises and extracts key information from over five years of research conducted by the iManage team, led by Professor Jeremias Adams-Prassl. Our analysis focuses on the UK and EU contexts, but may be applicable to other jurisdictions with similar legal landscapes or be drawn upon for comparative analysis. The toolkit only offers a snapshot of our research, with the full published papers linked where applicable and a further research list at the bottom of the page.
What existing levers in current legal and regulatory frameworks can be used to protect workers from the harms of algorithmic management?
The use of technology to evaluate and manage workers is not a new phenomenon. Frederick Taylor was already arguing for a scientific management approach to labour by the late 1800s, seeking to maximise efficiency by reducing the time taken for individual tasks. What is new is the scale, ubiquity and impact of these practices. – Aislinn Kelly-Lyth and Anna Thomas
Involvement in decisions about algorithmic management: information and consultation rights
Information and consultation rights refer to the legal requirements for employers to provide information to employees and their representatives, as well as duties to consult with these groups with a view to reaching agreement. In the development of algorithmic management systems and their subsequent deployment in workplaces, information and consultation rights can serve as powerful provisions to foster effective social dialogue and ensure worker voice in technology design.
The EU Platform Work Directive (PWD) develops the requirements in the EU Information and Consultation Directive, including:
- Where decisions are likely to lead to the introduction of or substantial changes in the use of automated monitoring and decision-making systems, Member States must ensure ‘information and consultation of platform workers’ representatives or, where there are no such representatives, of the platform workers concerned by digital labour platforms' (Article 13(2) PWD)
- Note that this does not extend to ‘persons performing platform work’ who are not ‘platform workers’—legitimately self-employed persons performing platform work (Recital 54 PWD; read more about the scope of the PWD here*)
- The primary negotiating counterparty for the employer is expected to be employees’ representatives (Article 13 PWD)
- Worker representatives can expect that information is ‘given in due time’ before consultation in light of the complexities of algorithmic management (Recital 52 PWD)
- The information and consultation rights irrespective of whether the systems are developed in-house, or by an external third party data processor on behalf of the digital labour platform (Recital 55 PWD)
- If digital labour platforms have 250+ workers, the costs of an expert can be paid for by the platform to offer assistance to worker representatives (Article 13(3) PWD)
*Note, this paper comments on the proposal for the PWD but its analysis on scope still applies for the PWD that is now in force.
Employers who are undertaking a Data Protection Impact Assessment (DPIA) can be required to consult workers or their representatives ‘where appropriate’ :
- Unions can consider where and how to argue that it is necessary to consult with worker representatives prior to deployment of new algorithmic management systems, or significant changes to existing systems, as these have been well-demonstrated to pose risks to worker health and safety and other substantive rights
- The logics or algorithms or processes used to make significant decisions about workers (e.g., in recruiting, work organisation, performance evaluation, discipline, etc.) are typically not going to be genuinely competitively sensitive, so arguments that employers (or technology providers) don’t need to provide explanations or access to data on grounds of trade secrets or intellectual property should be scrutinised.
- Where the security of processing operations (or commercial interests) is a genuine concern, this cannot result in the complete failure to consult or provide information (explanations/access to data) (see Recital 63 EU/UK GDPR; Dun & Bradstreet). Where necessary, the conflict should be adjudicated by a competent supervisory authority (see Dun & Bradstreet).
- There are more stringent requirements for employers to consult the national supervisory authority prior to processing if the algorithmic management system would pose a ‘high risk’ to the data subjects (e.g., use of new technologies, like surveillance systems) which the employer cannot sufficiently address (Article 36 UK GDPR; Article 36 EU GDPR)
Transparency into algorithmic management practices: data access rights
Without transparency into the workings of algorithmic management and their deployment in workplaces, trade unions cannot understand how workers’ rights and conditions will be affected. A lack of transparency also poses evidential difficulties to workers and their representatives when challenging or litigating against employers or algorithmic management developers.
The EU Platform Work Directive (PWD) offers both workers and their representatives transparency over algorithmic management. Indeed, Article 1(c) of the PWD even states that its purpose is to improve transparency in platform work, and Article 1(b) explicitly mentions this in the context of algorithmic management.
- Digital labour platforms are required to declare work performed by platform workers to the competent authorities of the Member State where the work is performed (Article 9(1) PWD)
- Workers, their representatives and national competent authorities shall receive information about the use and function of automated decision-making systems, including: the main parameters of the systems and their relative importance; the grounds for decisions that restrict, suspend or terminate platform work; and categories of decisions taken or supported by the system (Article 9(1)(b) PWD)
- Workers, their representatives and national competent authorities shall receive information about the use and function of automated monitoring systems, including the categories of data and actions monitored and evaluated by the systems; the aim of monitoring; how the system carries out monitoring (Article 9(1)(a) PWD)
- Platform workers should also receive information about the algorithmic management system (at the latest) on their first working day, prior to the introduction of changes which affect their working conditions, and at any time upon request (Article 9(3) PWD)
- Competent authorities and worker representatives are entitled to gain information (that is updated every 6 months) about algorithmic management systems, including: terms and conditions of the digital labour platform, employment status and number of platform workers engaged, working activity of the platform workers (e.g., duration of activity, number of hours worked etc) (Article 17(1) PWD; Article 17(4) PWD)
- Competent authorities and worker representatives can request additional clarifications on information provided by the digital labour platforms (Article 17(5) PWD)
Since the data protection regime seeks to protect data subjects in the context of personal data processing, its rights are limited and individualist in nature (read more about the limitations here). However, certain provisions can be used en masse to provide greater systemic clarity over workplace practices (read more here). Data Subject Access Requests (DSARs) can be revealing, specifically for individual decisions that are outputted by algorithmic management systems. In this context, 'data subjects' refer to workers.
- Data subjects have an ex post right of access to data about the ‘logic involved’ in automated decision-making which has legal or similarly significant effects on them (Article 15(1)(h) UK GDPR; Article 15(1)(h) EU GDPR)
- The right to be informed provides workers with information about how their personal data is collected and used by algorithmic management systems, e.g., what data is being processed and who it will be shared with (Article 13 UK GDPR; Article 13 EU GDPR)
- The information provided by data controllers (e.g., employers) about data processing must be concise, transparent, intelligible, easily accessible, and it must use clear and plain language (Article 13 UK GDPR; Article 12 EU GDPR)
- Data subjects can also request a copy of the personal data held by their employers (Article 15(3) UK GDPR; Article 15(3) EU GDPR)
Looking for further information?
Find a list of the research the iManage team conducted and drew on to develop the employee toolkit.
Algorithmic management:
The following article by Aislinn Kelly-Lynth and Anna Thomas and provides an overview of the impacts of algorithmic management. This Wired article also provides a shorter, lay summary of algorithmic management and the work of the iManage team.
Data protection law:
Between 2022 and 2024, the iManage team drew together resources for Data Protection Day curated around workers as data subjects:
- The Worker as a Data Subject: Resources for Data Protection Day (2022)
- Workplace Data Processing: Resources for Data Protection Day (2023)
- Data Processing at Work: Resources for Data Protection Day (2024)
In the following article, Regulating algorithmic employment decisions through data protection law, Halefom Abraha examines the extent to which the GDPR offers the necessary tools to protect workers from harm stemming from algorithmic management. In another piece, Halefom Abraha digs into Article 88 and EU Member State Employee Data Protection rules following the CJEU's decision in Hauptpersonalrat der Lehrerinnen concerning the live streaming of video classes in Germany during COVID. An additional piece by Halefom Abraha also focuses on Article 88 in upholding privacy in the workplace which is described as a 'doubled-edged sword'.
Platform Workers Directive:
For an analysis of the Platform Workers Directive with respect to algorithmic management, see Fortifying the algorithmic management provisions in the proposed Platform Work Directive by Michael Veale, Michael 'Six' Silberman, and Reuben Binns. A summary of the Platform Workers Directive can be found on blogs here and here, alongside blog posts (here and here and here) which criticises the limited scope of the legislation. Note, this research was conducted while the Platform Workers Directive was being debated, so its analysis is focused on the proposal/draft version of the Directive.
EU AI Act:
For an overview of the applicability of the EU AI Act to algorithmic management, see Aislinn Kelly-Lyth's article in the Comparative Labor Law & Policy Journal.
This site is part of a project that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant agreement No. 947806).
If you have any questions or feedback about this site or our research, please get in touch by emailing ai.work@law.ox.ac.uk.