en.Wedoany.com Reported - The European Union's recruitment office, facing capacity constraints due to a surge in job applications, is planning to introduce an AI tool based on US technology to screen and rank candidates, aiming to improve hiring efficiency.
This AI transformation of the EU's recruitment system comes against a backdrop of technical issues plaguing its long-standing efforts to update the hiring process through an online examination system, which has drawn criticism of the European Personnel Selection Office (EPSO). After two years of development, an AI tool named the "Job Matching Application" has neared its final testing phase. According to a senior European Commission official, if testing goes smoothly, it could be made available to a limited group of users as early as this summer.
The latest round of the EU's general competition (AD5) attracted 174,922 applications, nearly three times the expected number. Successful candidates will be placed on a reserve list for permanent civil servant positions. EU officials anticipate that future recruitment competitions will continue to generate similarly large pools of candidates, requiring recruiters to sift through tens of thousands of CVs while maintaining consistent standards. One EU official stated, "We simply cannot continue processing the new lists of successful candidates generated by EPSO; it's like finding a needle in a haystack."
The AI tool has been developed by Accenture based on the large language model (LLM) of the US AI company Anthropic and is hosted on Amazon Web Services (AWS). The system will be used to identify, score, and rank candidates within the EU recruitment process. In addition to its application for EPSO exams, it is also planned for filling internal vacancies and streamlining the recruitment of temporary and contract staff. According to the senior Commission official, the current recruitment process takes an average of three to four months.
Under the new system, applicants will submit their CVs and supporting documents through the EU's single candidate portal. According to internal documents seen by Euractiv, these documents will be pooled into a centralized talent database called the "Unified Talent Pool." Recruiters will then be able to search, filter, and rank candidates using a combination of traditional screening, AI-assisted scoring, and semantic matching. Semantic matching aims to understand how well a candidate's experience aligns with job requirements. The European Commission believes the new system will help identify suitable candidates who might be missed by traditional searches, while also alleviating the administrative burden caused by the growing scale of EPSO exams.
The technology will serve only as an auxiliary tool; recruiters will define the screening criteria, and selection panels retain the final decision-making authority. When the tool is used, candidates will be informed via an AI disclaimer in the job advertisement. Candidates will also have the right to request human intervention before an interview and can challenge any decision.
The project has also raised concerns about algorithmic bias and the risks associated with using candidate information. Internal communications among EU officials seen by Euractiv reveal concerns that applicant information could be used for purposes beyond those intended. Furthermore, issues such as the opacity of decision-making processes related to large language models, information fabrication, and their suitability for recruitment have also drawn attention. One study found that LLMs may show a preference for CVs generated by LLMs over those written by humans or by other AI models not trained for language processing and generation.
The project also raises questions regarding EU technological sovereignty and privacy, at a time when the EU is seeking to reduce its dependence on foreign technology platforms. A potential privacy issue discussed by both parties concerns the source of CVs used to train and test the tool's technology. In response, a senior European Commission official stated that the tool "was developed and tested using only anonymized and fictitious CVs," with testing using real CVs only beginning recently. According to internal European Commission documents, candidate data will not be used to train Anthropic's AI model, and their personal information will not be shared with the LLM provider. The European Data Protection Supervisor (EDPS) is working with the European Commission to assess whether the tool complies with data protection rules.
Recruitment is one of the most strategically important functions of the EU's administrative system, yet the technology on which this process relies is now largely developed outside its borders. The project reveals the trade-offs faced by EU institutions in modernizing their own administration: a system designed to make hiring faster and more consistent must now prove it can also meet the standards of transparency, privacy, and independence demanded by Brussels.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
