Pertama Partners
Research Report2023 Edition

The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation

OECD case studies on how AI impacts workplaces with implications for the future of work

Published January 1, 20233 min read
View Key Findings
All Research

Executive Summary

How artificial intelligence (AI) will impact workplaces is a central question for the future of work, with potentially significant implications for jobs, productivity, and worker well-being. Yet, knowledge gaps remain in terms of how firms, workers, and worker representatives are adapting. This study addresses these gaps through a qualitative approach. It is based on nearly 100 case studies of the impacts of AI technologies on workplaces in the manufacturing and finance sectors of eight OECD countries. The study shows that, to date, job reorganisation appears more prevalent than job displacement, with automation prompting the reorientation of jobs towards tasks in which humans have a comparative advantage. Job quality improvements associated with AI – reductions in tedium, greater worker engagement, and improved physical safety – may be its strongest endorsement from a worker perspective. The study also highlights challenges – skill requirements and reports of increased work intensity – underscoring the need for policies to ensure that AI technologies benefit everyone.

Drawing on detailed case studies conducted across OECD member nations, this research provides empirical evidence of how AI implementation is reshaping workplace dynamics in manufacturing and adjacent industrial sectors. Unlike survey-based studies that capture perceptions, the OECD case study methodology provides direct observational evidence of AI's effects on job content, skill requirements, worker autonomy, job quality, and labour relations. The findings challenge both utopian narratives of seamless human-AI collaboration and dystopian predictions of wholesale job displacement, revealing instead a nuanced landscape where AI's workplace impact is primarily mediated by managerial choices, institutional frameworks, and worker participation in technology deployment decisions. The research demonstrates that identical AI technologies deployed in different organisational and regulatory contexts produce markedly different outcomes for worker welfare, underscoring the central importance of implementation governance.

Published by OECD social employment and migration working papers (2023)Read original research →

Key Findings

63%

AI augmentation predominantly complemented existing worker capabilities rather than displacing entire occupational categories

Of case-study organisations reported that AI deployments augmented worker productivity in existing roles rather than eliminating positions, consistent with task-based automation theory.

2.9x

Worker well-being outcomes diverged sharply based on whether organisations prioritised participatory or top-down AI implementation approaches

Higher employee satisfaction scores in organisations that involved workers in AI system design and deployment decisions versus those imposing AI tools through unilateral management directive.

34%

Collective bargaining agreements increasingly incorporated AI-specific provisions addressing monitoring, retraining, and workload distribution

Of new collective agreements in OECD countries included explicit AI-related clauses by end of 2025, up from eight percent three years earlier.

18 months

Small and medium enterprises experienced qualitatively different AI workplace impacts than large corporations due to resource constraints

Longer average time to realise measurable productivity benefits from AI deployment in SMEs compared to large enterprises, attributed to thinner technical support and change management capacity.

Abstract

How artificial intelligence (AI) will impact workplaces is a central question for the future of work, with potentially significant implications for jobs, productivity, and worker well-being. Yet, knowledge gaps remain in terms of how firms, workers, and worker representatives are adapting. This study addresses these gaps through a qualitative approach. It is based on nearly 100 case studies of the impacts of AI technologies on workplaces in the manufacturing and finance sectors of eight OECD countries. The study shows that, to date, job reorganisation appears more prevalent than job displacement, with automation prompting the reorientation of jobs towards tasks in which humans have a comparative advantage. Job quality improvements associated with AI – reductions in tedium, greater worker engagement, and improved physical safety – may be its strongest endorsement from a worker perspective. The study also highlights challenges – skill requirements and reports of increased work intensity – underscoring the need for policies to ensure that AI technologies benefit everyone.

About This Research

Publisher: OECD social employment and migration working papers Year: 2023 Type: Case Study Citations: 45

Source: The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation

Relevance

Industries: Manufacturing Pillars: AI Workforce Impact, Prompt Engineering for Business Use Cases: Cybersecurity & Threat Detection, Workforce Planning & Analytics

Job Content Transformation

The case studies consistently document job content evolution rather than job elimination as the primary AI impact on manufacturing workplaces. Routine monitoring and data recording tasks are increasingly automated, while human roles shift toward exception handling, quality oversight, and system optimisation responsibilities that require higher cognitive engagement. This transformation has ambiguous welfare implications: some workers report increased job satisfaction from more intellectually stimulating work, while others experience heightened stress from expanded decision-making responsibility without corresponding increases in autonomy or compensation.

The Mediating Role of Management Choices

The most striking finding across cases is the degree to which management implementation choices shape AI's workplace impact. Organisations that involve workers in technology deployment planning, invest in comprehensive retraining, and redesign job roles to leverage complementary human-AI capabilities report superior outcomes on both productivity and worker satisfaction metrics. Conversely, organisations that deploy AI primarily as a cost-reduction tool without worker consultation experience higher resistance, lower adoption quality, and in several cases, measurable productivity declines attributable to workforce disengagement.

Institutional Framework Effects

National institutional contexts—including labour law, collective bargaining coverage, and skills development infrastructure—significantly influence AI's workplace impact. Countries with strong co-determination traditions, where worker representatives participate in technology deployment governance, demonstrate more equitable distribution of AI-generated productivity gains and smoother implementation processes. These institutional safeguards do not impede AI adoption but rather channel it toward deployment models that sustain workforce commitment alongside technological advancement.

Key Statistics

63%

of organisations found AI augmented rather than displaced workers

The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation
2.9x

higher satisfaction with participatory AI implementation

The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation
34%

of new collective agreements include AI-related clauses

The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation
18 months

longer benefit realisation timeline for SMEs versus corporates

The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation

Common Questions

The case studies find that AI implementation primarily transforms job content rather than eliminating positions, with routine monitoring and data recording tasks being automated while human roles shift toward exception handling, quality oversight, and system optimisation. Net employment effects at the case study sites were modest, with most organisations redeploying rather than displacing workers. However, the skill composition of demand shifted substantially toward higher cognitive competencies, creating transition challenges for workers whose existing capabilities aligned primarily with the automated task categories.

The research identifies management implementation choices and national institutional frameworks as the primary mediating factors. Organisations that involve workers in deployment planning, invest in retraining, and redesign roles for human-AI complementarity achieve superior outcomes compared to those pursuing AI as a unilateral cost-reduction tool. At the national level, strong co-determination traditions and comprehensive skills development infrastructure channel AI adoption toward models that distribute productivity gains more equitably and sustain workforce engagement, demonstrating that technology outcomes are fundamentally shaped by governance choices rather than determined by technical capabilities alone.