Title page
Contents
Résumé 4
Abstract 5
Acknowledgements 6
Executive summary 9
Synthèse 11
Zusammenfassung 14
1. Introduction: Objectives and definitions 17
1.1. Introduction and objectives 17
1.2. Definitions: Categories of AI and the use of Big Data 18
2. Opportunities and challenges for PES in adopting AI 22
2.1. Key opportunities for PES in adopting AI 23
2.2. Central challenges and risks faced by PES in utilising AI 27
3. Mapping AI use in PES 31
3.1. High-level overview of AI use in PES across the OECD 31
3.2. AI use to understand jobseeker needs and provide targeted support 33
3.3. AI-powered labour market matching and employer services 44
3.4. AI tools to assist PES administrative activities and knowledge generation 51
4. Impact of AI adoption on PES staff 55
4.1. AI adoption can significantly impact the day-to-day work of staff and PES workforce needs 55
4.2. PES must be proactive in addressing the challenges AI adoption may bring for staff, including resistance and lack of trust 57
5. Conclusion and key considerations 59
References 62
Annex A. Overview of AI use by PES 71
Annex B. Technical Annex 72
Table 3.1. While traditional rule-based chatbots remain prominent, several PES are engaging in AI-driven applications 34
Table 3.2. Seven OECD PES currently have AI-driven profiling solutions in place 39
Table 3.3. Several PES are using AI to assist career management and job-search orientation 42
Table 3.4. Eight PES in OECD countries have implemented AI-driven vacancy matching solutions 45
Table 3.5. Several PES are using additional AI tools for employers to further support job matching 48
Figure 1.1. Under the umbrella of AI, various sub-categories of methods exist 20
Figure 2.1. Adopting AI presents a number of opportunities and challenges for PES 22
Figure 2.2. In contrast to black box models, explainable AI models enable better transparency and explainability 28
Figure 3.1. AI presents various opportunities for PES across central areas of activity 31
Figure 3.2. Half of PES have implemented AI initiatives 32
Figure 3.3. AI is being used across all key areas of PES activities 33
Figure 3.4. The job-matching tool of the Korean PES conducts matching across two dimensions 46
Figure 4.1. More widely, technological adoption has significantly impacted PES staff 55
Figure 4.2. Most PES take steps to support staff in adapting to new technologies 57
Boxes
Box 1.1. The recent boom in generative AI and large language models (LLMs) may yield significant benefits for PES 21
Box 2.1. A workplace study showed that access to a generative AI tool boosted worker productivity, particularly for low-skilled workers 26
Box 3.1. User-centric smart bots can help break down silos between public organisations 37
Box 3.2. Estonia's automated benefit system has provided efficiencies for both staff and clients 52
Annex Tables
Table A A.1. AI use by PES by country/region and area of PES activity 71
Table A B.1. Performance metrics of profiling tools 74
Annex Figures
Figure A B.1. Key determinants of model performance 75