Privacy International's proposed amendments to the EU Platform Work Directive
PI submitted amendments to the EU's Directive on Working Conditions and Platform Work to raise concerns about the threats workers face when they are subjected to automated decision-making systems. This is part of our work defending workers' rights to access information as well appropriate mechanisms to defend their interests, and, more broadly our efforts to ensure that automated decision-making systems are subjected to effective and robust scrutiny.
Key points
- The current draft of the Directive on Working Conditions and Platform Work contains shortcomings with regard to general principles relating to the fundamental right to privacy, including privacy by design and by default, transparency and decisions affecting platform workers’ working conditions
- We focused on the threats to workers’ rights that can materialise at three key levels: when personal data is being collected, when decision-making occurs with no transparency and when decision-making systems operate without human oversight.
- PI's Managed by Bots campaign exposed how such threats can have can be financially and emotionally devastating for workers and urges the Member of the European Parliament to consider these amendments
Advocacy
Post date
7th October 2022
UPDATES:
- The Platform Work Directive (PWD) became law in October 2024. Member States have until December 2026 to implement its provisions into their national law. The final text of the PWD provides the following:
- A person performing platform work cannot be fired or suspended based on a decision made by an algorithm or automated decision-making systems.
- Platforms must provide upfront information, in "clear and plain language", about their use of automated decision-making systems.
- Platform workers have the right to an explanation of the reasons for automated decisions "without undue delay". Certain essential decisions must always be accompanied by a written explanation by the time such decision takes effect.
- Platforms must provide for human review of automated decisions by people with the appropriate levels of resourcing, competence and authority.
- Digital labour platforms may not process certain types of personal data of platform workers, such as data on their emotional or psychological state and personal beliefs. Nor can they collect data while someone is not at work.
- The Directive also creates a presumption that platform workers are employees (ie rather than self-employed), based on real-world fact rather than legal fiction. The burden of proof lies with the platform to prove there is no employment relationship.
- The adoption of the PWD is an important step forward, but the text still falls short of some of our advocacy demands. For instance, the algorithmic management provisions apply to all platform workers regardless of employment status, but the reality is that those recognised as employees are more likely to be able to systematically make use of them.
- The PWD also creates an unusual situation in which platform workers have access (on paper) to workplace protections not available to other workers. In June 2025, the European Parliament’s Committee on Employment and Social Affairs published a draft report recommending the European Commission to initiate the legislative process for an EU Directive on algorithmic management in the workplace more widely, following the publication of a Commission study earlier that year.
- PI is also now pushing for the International Labour Organization to put in place a global and legally binding Convention to protect platform workers from the harms of algorthmic abuse.
Algorithmic management fundamentally relies on the availability of data to make decisions. The impact that these decisions can have on workers can be financially and emotionally devastating.
PI has previously exposed this issue through the Managed by Bots campaign - in which we called for the conditions under which data is collected and processed to be subjected to effective and robust scrutiny.
To address the threats to workers' privacy and prevent the ascent of inscrutable black box algorithms, employers must be accountable and transparent in relation to their data collection and processing practices. Platform workers must have access to mechanisms which enable data security and provide them with information in way that empowers them and ensures that their rights are protected at least as much as employees and workers working under ‘traditional’ employment contracts.
While we welcome the Platform Directive as a mechanism to protect workers’ rights in response to transformations in the workplace, we note that the proposal put forward by the European Commission contains certain shortcomings with regard to general principles relating to the fundamental right to privacy, including privacy by design and by default, transparency and decisions affecting platform workers’ working conditions, which are all detailed in the report below. It is crucial that these are effectively addressed by the European Parliament through the introduction of specific amendments to ensure that the aim of the Directive is not undermined, and that workers’ rights remain protected.
The three key principles defended in this report are summarised in essence by the following:
- The data produced by workers’ devices can be extremely revealing. Accessing this data is intrusive and threatens their enjoyment of their fundamental rights. To limit threats to workers’ privacy, only data which is necessary for the performance or attribution of work should be collected and only during periods of time agreed by workers.
- Decision-making systems are complex and fallible systems with which real human beings interact and depend on. To ensure that workers, users, researchers, and the general public can assess, understand, and challenge these systems, transparency about the parameters considered by these systems and their functionality is key, in particular when such systems are relied on to make significant decisions such as account termination.
- Given the fallible nature of decision-making algorithms, decisions affecting platform workers’ working conditions should never be taken without human review. The aim of human review is to ensure that both the parameters and logic deployed are correct.
