The European Union AI Office has initiated a recruitment drive for 40 contractual agents, marking the transition from establishing a regulatory framework to building an active inspection regime. By adding these roles—technology specialists, legal officers, operations specialists, and paralegals—the office is scaling its capacity to handle the practical demands of the AI Act. This expansion brings the total staff to approximately 165, a modest figure when measured against the breadth of the enforcement scope that will unfold through 2028.
This operational shift follows the August 2, 2026, commencement of enforcement, where the AI Office and national authorities began deploying tools such as requests for information (RFIs), model evaluations, requests for access (RFAs), and physical inspections of provider premises, as detailed in the Aug 2 enforcement press release. The infrastructure now includes three distinct complaint channels: a general tool for natural and legal persons, a whistleblower mechanism for industry professionals, and a specific channel for downstream providers. These tools are designed to feed directly into an enforcement process backed by significant sanctioning powers, including penalties of up to 35 million euros or 7 percent of worldwide turnover for prohibited practices.
The specific composition of these 40 new hires will determine the character of the office’s oversight. A heavy reliance on technology specialists suggests a code-first approach to General Purpose AI (GPAI) audits, while a focus on legal officers and paralegals points toward a procedure-first methodology centered on compliance documentation and administrative accountability. Because the AI Office has positioned this recruitment alongside its enforcement framework and complaint tools, it is clear that these agents are intended to serve as the front line for the ongoing buildout of the regulatory apparatus.
The European approach stands in contrast to the strategies adopted by the United States and China. The U.S. federal government, guided by Executive Order 14409, has rejected mandatory licensing or preclearance in favor of an enforcement-first posture that relies on existing criminal statutes like the Computer Fraud and Abuse Act (CFAA). There is no comparable federal agent-specific governance framework in the U.S., nor are there centralized registration or safety requirements, a topic explored in the Federal Agent Regulation Gap article. Conversely, China has implemented a structured registration system for public-facing generative AI, including a national framework for intelligent agents that mandates three-tier decision authorization and high-risk sector filings.
For AI developers operating globally, the EU’s move represents a shift toward a more granular, interventionist style of oversight. The enforcement timeline is staggered, beginning with GPAI obligations and transparency requirements, moving to prohibitions on intimate content and CSAM in December 2026, and extending to high-risk AI systems under Annex III by December 2027, with regulated products following in August 2028. The Code of Practice on transparency, which has garnered over 180 signatories, currently covers marking and labeling but notably excludes agent and chatbot disclosure requirements, leaving a gap that future enforcement actions may eventually address.
Despite the clarity of the recruitment goals, significant uncertainty remains regarding the efficacy of this staffing level. With a total headcount of roughly 165, the AI Office must manage a vast and technically complex landscape of providers and systems. The success of this enforcement regime will depend not only on the number of agents hired but on their ability to integrate technical evaluation with the legal rigor necessary to sustain the enforcement actions the Commission has now authorized.