EU REGULATORY UPDATE | EU AI Act obligations applying from 2 August 2025
From 2 August 2025, further provisions of the EU AI Act (Regulation 2024/1689) became applicable, including parts of the framework for general-purpose AI models. This marked another step in the phased implementation of Regulation (EU) 2024/1689.
General-purpose systems may be embedded in CV analysis, drafting of evaluations, employee help desks and management analytics. The employment risk depends not only on the model but on the way the employer deploys it and the decision that follows.
Foreign groups may allow employees to use a global AI assistant while local managers feed it personal or sensitive information. That can create data-protection, confidentiality and discrimination issues even before the system qualifies as a high-risk employment application.
Employers should define approved and prohibited workplace uses, prevent sensitive data from being entered without a lawful basis, train decision-makers and review vendor documentation. Generative output should assist a documented human process, not become an unrecorded reason for an employment decision.