Part 1 (read here: The EU Platform Work Directive: What it is and why it matters) outlined the background to the Directive and clarified which employers fall within scope.

In this second instalment of our three-part series on the EU Platform Work Directive (Directive (EU) 2024/2831), we examine in detail the legislative obligations that will transform the regulatory landscape for digital labour reaching obligations.  This Part 2 focuses on three core pillars of legal change: employment status, algorithmic management, and data protection.  Each of these pillars introduces complex and far-reaching obligations.

Part 3 will provide a practical compliance roadmap for Irish employers preparing for the 2026 transposition deadline. 

The rebuttable presumption of employment 

The Directive introduces a rebuttable presumption of employment, a mechanism specifically designed to address the misclassification of platform workers across the EU. Where indicators of “direction and control” exist, the individual will be presumed to be an employee rather than self-employed.

The Directive requires Member States to define procedures for making these determinations, referencing “national law, collective agreements and established case law.”  

In October 2023, the Supreme Court’s judgment in The Revenue Commissioners v Karshan (Midlands) Ltd t/a Domino’s Pizza clarified the multifactor test for determining employment status, emphasising the reality of the working relationship over contractual labels and setting out a structured five question analytical framework. Following the Karshan judgment, a Code of Practice issued by Revenue, the Department of Social Protection, and the Workplace Relations Commission provided a framework for determining whether a worker is an employee or self‑employed. It adopts the Court’s five‑step test, emphasising that the real working relationship (not a contractual label) determines status, and introduces three initial “filter questions” on remuneration, personal service, and control to assess if an employment contract is even possible.  

The Karshan decision and the Code will become even more relevant once the Directive becomes operative. The presumption will not replace Ireland’s existing framework but it will reinforce the direction of travel already established. 

Under the new regime, the platform, not the worker, must disprove the existence of an employment relationship. For employers, this will mean proactively keeping evidence to rebut the presumption and being prepared for documentation to come under scrutiny, including app architecture, contractual arrangements and worker communication logs etc. Employers should also prepare for a potential increase in misclassification claims once the presumption becomes operative.  

Algorithmic management 

The Directive introduces the EU’s first dedicated regulatory framework for algorithmic management, reflecting concerns that opaque automated systems exercise significant control over working conditions without transparency or human oversight.  

Platforms must inform workers, no later than the first working day, about any use of: 

• automated task allocation; 

• automated performance evaluation; 

• automated or semiautomated decision-making; 

• monitoring technologies that track behaviour, location, or productivity. 

These notifications must also be issued whenever systems are updated or modified in a way that could affect working conditions.  

In this respect, the Directive is more onerous than the EU AI Act, which focuses on risk management but does not impose worker-specific disclosure requirements.

The Directive mandates that automated decisions affecting working conditions must be subject to: 

• review by qualified staff; 

• authority to intervene, including suspending or overturning automated decisions; and 

• regular competence training, ensuring reviewers understand system logic and limitations.  

This requirement is a direct legislative response to cases where workers were suspended or “deactivated” without meaningful explanation; these situations will now be deemed incompatible with EU labour protections. 

Every two years, platforms must conduct evaluations of: 

• how algorithms affect equal treatment; 

• whether automated decisions introduce discriminatory patterns; and 

• whether earnings, access to work, or task distribution are skewed by algorithmic bias.  

These biannual assessments will impose significant technical and administrative burdens, including the need for audit-ready data logs and explainability documentation.

Data protection and worker privacy

The Directive also significantly tightens the data protection regime applicable to platform work. 

The Directive declares that worker consent is no longer a valid legal basis for most processing activities in platform work, because consent in hierarchical or economically dependent contexts is not freely given.  

This means platforms must rely on alternative legal grounds, most likely “performance of a contract” or “legitimate interest”, but such grounds will be strictly interpreted given the Directive’s protective purpose. 

The Directive prohibits the automated processing of: 

• emotional or psychological state data; 

• private communications between workers; 

• personal information irrelevant to platform work; and 

• data used to anticipate or predict exercise of fundamental rights (e.g., union activity).  

This prohibition will require many platforms to reconfigure monitoring tools, especially those that rely on behavioural analytics or sentiment analysis.  

Conclusion 

Each of the pillars outlined in this Part 2 introduces complex obligations that Irish employers must now prepare for. In Part 3, we will translate this detailed legislative analysis into a practical compliance blueprint, including audits, risk assessments, and workforce management.