The EU employment law of 2026 shifted the responsibility of proof onto you.

      If your HR team spent this year focusing on legislation instead of recruitment, that is not merely an issue of scheduling. Three new pieces of European employment law will take effect in 2026, with a fourth set to arrive in December. They reflect a fundamental design principle that many companies have yet to recognize.

      The central theme across all of these laws is the shifting burden of proof. In the past, if a worker suspected misclassification, unequal pay, or an unjust algorithmic decision, they had to provide proof. Under the new rules being introduced, it is increasingly the company's responsibility to demonstrate compliance. Compliance is evolving from a simple policy matter to an evidence challenge.

      On June 7, the EU Pay Transparency Directive required all 27 member states to implement national legislation by that date, with only four successfully doing so: Italy, Slovakia, Lithuania, and Malta. Greece followed soon after, while the rest remain somewhere between partial implementation and no draft law published at all, including Germany, Spain, and Sweden; France, the Netherlands, and Denmark aim for January 1, 2027. Sweden has completely halted its transposition efforts while renegotiating.

      The European Commission did not extend the deadline and has indicated that infringement actions under Article 258 TFEU may be forthcoming. Workers in states that did not comply may also pursue damages claims against their governments based on the Francovich ruling. This situation puts employers in a passive role, observing from the sidelines.

      The directive itself is straightforward: employers must share pay ranges with candidates before interviews, inquiries about salary history are prohibited, and employees can request pay data comparisons based on gender. Gender pay gap reporting is set to commence, with initial reports due in June 2027.

      For businesses recruiting across multiple EU markets, the challenge lies not in the rules themselves but in the fact that a singular job advertisement may be permissible in one member state but not in another, determined solely by the employer's location.

      August 2 marked the first hard deadline for obligations under the EU AI Act. Article 50’s transparency requirements began on that date, imposing fines of up to €15 million or 3% of global annual revenue, regardless of whether a system is categorized as high risk. This affects more HR technologies than many teams recognize—such as chatbots for candidate questions, AI-generated recruitment messages, and any interaction where individuals reasonably expect to deal with a machine.

      Initially, stronger obligations for high-risk recruitment and HR tools were scheduled for this month, but the AI Omnibus has postponed them to December 2, 2027. This delay is a temporary relief rather than a full reversal, and it brings an unintended consequence. Numerous companies had structured governance timelines based on an August 2026 deadline, and with its alteration, many quietly halted related projects.

      On June 16, the Netherlands exemplified this shift when its senate passed bill 36.783, amending Book 7 of the Civil Code to create a legal presumption of employment for work compensated under €38 per hour, based on a reference date of January 1, 2026. When a self-employed worker invokes this presumption, the hiring company must establish that no employment relationship exists. If they cannot, the worker gains employment protections, including continued sick pay and protection against dismissal, as outlined in a Royal Decree yet to set a start date.

      This change follows the resumption of enforcement in 2025 after a decade-long moratorium, retroactively correcting issues back to January 1, 2025, and reinstating significant penalties this year. Any remaining leniency will end on January 1, 2027. The Netherlands’ changes are more about timing than direction, as it is anticipated that several member states will also incorporate false self-employment enforcement into their platform work transpositions.

      The next critical deadline is December 2, when the EU Platform Work Directive must be transposed. This regulation introduces a rebuttable presumption of employment whenever direction and control are evident, with the burden resting on the platform to prove otherwise. However, attempting to challenge this presumption does not halt its applicability, and for existing relationships as of that date, it will apply moving forward without retroactive effect.

      Two notable points extend beyond the directive's title. The definition of a digital labor platform is broad enough to encompass freelance marketplaces, staffing platforms, and agency models, not just food delivery or ride-hailing services. Furthermore, the chapter on algorithmic management also affects genuinely self-employed individuals, not just those covered by the presumption.

      This chapter establishes the world's first binding regulations on algorithmic management, reaching beyond what the AI Act stipulates. Platforms are required to disclose how their automated systems make decisions regarding work distribution, pricing, and account limitations, and significant decisions must include human oversight. Strict limitations on data processing are imposed, banning the use of emotional or psychological data, health information, private conversations, and data collection when individuals are off duty.

      As with pay