Abstract
This dissertation examines the efficacy of reskilling programmes in protecting workers from artificial intelligence (AI)-related redundancy during periods of organisational hiring freezes. Through systematic literature synthesis, this study identifies the critical characteristics that distinguish genuinely protective reskilling initiatives from those offering merely superficial workforce development. The analysis reveals that proactive, employer-linked reskilling programmes targeting concrete AI-complementary roles demonstrate significantly higher retention rates—up to 64% greater than reactive training interventions. Effective programmes combine digital competencies with soft skills development, maintain tight linkages to real vacancies and redesigned roles, and operate within tripartite frameworks involving employers, government, and educational institutions. Conversely, stand-alone generic online courses demonstrate limited protective capacity, particularly for vulnerable worker populations. The findings highlight a substantial 42% adaptation gap where misaligned or inaccessible programmes fail to facilitate successful workforce transitions. This research contributes to labour economics and human resource management scholarship by establishing an evidence-based framework for evaluating reskilling programme effectiveness, offering practical guidance for policymakers and organisations navigating AI-driven workforce transformation.
Introduction
The rapid advancement and deployment of artificial intelligence technologies across global industries has precipitated profound transformations in labour markets, fundamentally reshaping employment landscapes and workforce requirements. Contemporary organisations face unprecedented challenges as AI systems increasingly automate tasks previously performed by human workers, creating both substantial displacement risks and emergent opportunities for workforce reconfiguration. This technological transition occurs against a backdrop of economic uncertainty, with many organisations implementing hiring freezes that constrain traditional pathways to employment whilst simultaneously accelerating automation investments to maintain competitive advantage.
The intersection of AI advancement and restrictive hiring practices creates a particularly precarious situation for workers whose roles face automation risk. Without access to new employment opportunities externally, these workers must rely upon internal reskilling pathways to maintain their employability and avoid redundancy. This context elevates the importance of understanding which reskilling approaches genuinely protect workers from displacement, as opposed to those that merely create an appearance of organisational investment in human capital without delivering substantive employment security.
The academic and practical significance of this inquiry extends across multiple domains. From a labour economics perspective, understanding effective reskilling mechanisms contributes to theories of human capital formation and technological unemployment. For human resource management scholarship, this research informs strategic workforce planning and development practices. From a policy perspective, identifying protective programme characteristics enables more effective allocation of public resources toward genuinely beneficial interventions. Socially, the consequences of inadequate reskilling are profound, potentially exacerbating inequality, diminishing social mobility, and undermining community stability in regions heavily affected by AI-driven displacement.
The World Economic Forum (2023) has projected that by 2027, approximately 83 million jobs will be displaced globally due to technological change, whilst 69 million new roles will emerge. This significant net displacement requires unprecedented attention to workforce transition mechanisms. Concurrently, the Organisation for Economic Co-operation and Development (2023) has emphasised that the quality and design of reskilling programmes substantially influence their effectiveness, suggesting that not all training investments yield equivalent returns in employment protection.
This dissertation addresses a critical gap in the extant literature by systematically examining the characteristics that distinguish genuinely protective reskilling programmes from those that fail to prevent AI-related redundancy. By synthesising evidence across multiple sectors, geographic contexts, and programme types, this research establishes an empirically grounded framework for evaluating and designing reskilling interventions during periods of constrained hiring.
Aim and objectives
The primary aim of this dissertation is to identify and critically evaluate the characteristics of reskilling programmes that genuinely protect workers from AI-related redundancy, particularly within organisational contexts characterised by hiring freezes.
To achieve this aim, the following objectives guide this research:
1. To examine the timing dimension of reskilling interventions, comparing the protective efficacy of proactive versus reactive training approaches in relation to workforce retention outcomes.
2. To analyse the importance of employer linkage and vacancy integration in reskilling programme design, evaluating how connections to specific emerging roles influence displacement prevention.
3. To investigate the relative effectiveness of different programme delivery models, including employer-led initiatives, public-private partnerships, and stand-alone educational offerings.
4. To identify the skill combinations—encompassing both digital competencies and soft skills—that demonstrate greatest protective value against AI-driven displacement.
5. To assess the role of support infrastructure and inclusive design in ensuring reskilling programme accessibility and effectiveness across diverse worker populations.
6. To develop an evidence-based framework for categorising reskilling programmes according to their protective capacity against AI-related redundancy.
Methodology
This dissertation employs a systematic literature synthesis methodology to address the research aim and objectives. This approach is particularly appropriate given the emerging and multi-disciplinary nature of the research topic, which spans labour economics, human resource management, educational theory, and technology policy. Literature synthesis enables the integration of diverse evidence streams to develop comprehensive understanding of complex phenomena whilst maintaining methodological rigour.
The literature search strategy encompassed multiple academic databases, including Scopus, Web of Science, and Google Scholar, supplemented by targeted searches of institutional repositories and grey literature from international organisations. Search terms combined variations of key concepts including “reskilling,” “upskilling,” “artificial intelligence,” “automation,” “workforce displacement,” “redundancy,” “training programmes,” and “employment protection.” Boolean operators facilitated systematic combination of search terms to maximise relevant retrieval whilst minimising irrelevant results.
Inclusion criteria specified English-language publications from 2019 to 2025, reflecting the contemporary relevance of AI-driven workforce transformation. Sources included peer-reviewed journal articles, conference proceedings, policy reports from governmental and intergovernmental organisations, and working papers from established research institutions. Quality assessment criteria prioritised empirical research, systematic reviews, and theoretically rigorous conceptual analyses, whilst excluding opinion pieces, blog posts, and sources lacking clear methodological foundations.
The analytical framework employed thematic analysis principles, with sources coded according to programme characteristics, outcome measures, sectoral contexts, and geographic scope. This enabled systematic comparison across studies and identification of convergent findings regarding protective programme features. Where quantitative evidence was available, effect sizes and comparative metrics were extracted to enable more precise characterisation of programme effectiveness differentials.
Limitations of this methodological approach include reliance upon published research, which may exhibit publication bias toward positive findings, and the heterogeneity of outcome measures across studies, which constrains direct quantitative comparison. Nevertheless, the breadth of sources examined and the consistency of findings across diverse contexts strengthen confidence in the conclusions drawn.
Literature review
### The AI employment disruption context
The proliferation of artificial intelligence across economic sectors has generated substantial academic and policy attention regarding workforce implications. Contemporary AI systems demonstrate capabilities extending well beyond routine task automation, increasingly performing cognitive functions previously considered exclusively human domains. This expansion of automation potential has heightened displacement concerns across occupational categories, including professional and knowledge work sectors traditionally considered insulated from technological unemployment.
Research by Kanagarla (2024) provides multi-sector analysis of AI-driven workforce disruption, documenting significant displacement patterns across manufacturing, services, administration, and logistics sectors. This work establishes the empirical foundation for understanding the scope and distribution of AI-related redundancy risk, demonstrating that displacement is neither uniform nor inevitable but shaped substantially by organisational and policy responses. The analysis reveals considerable variation in redundancy outcomes between organisations facing similar technological conditions, suggesting that intervention strategies significantly influence worker outcomes.
Maria et al. (2025) extend this analysis by examining the dual phenomena of job displacement and creation in AI-transformed labour markets. Their findings indicate that whilst automation eliminates certain role categories, it simultaneously generates demand for new positions in AI operations, data analytics, and human-AI collaboration functions. This displacement-creation dynamic underscores the importance of transition mechanisms that facilitate worker movement from declining to emerging occupational categories.
The service industry faces particular challenges, as documented by Kong, Li and Song (2025), who analyse AI-driven job replacement patterns and unemployment response strategies. Their research identifies the service sector’s combination of high automation potential and limited historical investment in technical training as factors exacerbating displacement vulnerability. These sectoral analyses collectively establish that AI-related redundancy risk varies substantially across industries, occupational categories, and organisational contexts, necessitating differentiated intervention approaches.
### Temporal dimensions of reskilling intervention
The timing of reskilling intervention relative to displacement events emerges as a critical determinant of programme effectiveness. Kanagarla (2024) provides compelling quantitative evidence on this dimension, finding that organisations implementing reskilling programmes proactively—before displacement becomes imminent—achieve 64% higher retention rates for at-risk workers compared to those initiating training only when redundancy threatens.
This substantial differential reflects multiple mechanisms. Proactive reskilling enables workers to develop competencies whilst still employed, maintaining income stability and psychological security that facilitate learning. Early intervention also permits more deliberate programme design aligned with emerging organisational requirements, rather than hastily constructed training responding to immediate crisis. Furthermore, proactive approaches signal organisational commitment to workforce development, potentially enhancing worker engagement and motivation during training.
The reactive training paradigm, conversely, typically operates under time pressure that constrains curriculum depth and pedagogical quality. Workers facing imminent redundancy may experience anxiety and uncertainty that impede learning effectiveness. Additionally, reactive programmes often lack clear employment pathways, as organisations implementing them may have already made strategic decisions reducing internal placement opportunities.
These temporal considerations carry significant implications for programme design and policy development. Chhibber, Rajkumar and Dassanayake (2025) emphasise the importance of anticipatory workforce planning that identifies roles facing future automation risk and initiates reskilling well in advance of technological implementation. This proactive orientation requires sophisticated labour market intelligence and organisational foresight capabilities that many employers currently lack.
### Employer linkage and vacancy integration
The connection between reskilling programmes and actual employment opportunities represents a second critical dimension distinguishing protective from ineffective interventions. Research consistently demonstrates that programmes maintaining tight linkages to real vacancies and redesigned roles achieve substantially better displacement prevention outcomes than those operating in isolation from employer requirements.
Depoo et al. (2025) examine AI implementation impacts on job transformation and competency requirements through case studies of Czech companies. Their findings document how organisations that integrated reskilling with role redesign achieved job transformation with minimal cancellations, as workers transitioned into modified positions rather than facing redundancy. This approach treats automation not as replacement of workers but as transformation of work, with training enabling workers to assume new responsibilities complementing AI capabilities.
Similar patterns emerge from automotive industry analysis by Costa, Machado and Conceição (2024), who document successful reskilling outcomes when training pipelines workers into specific emerging roles. Their case study demonstrates that clear articulation between training content and destination positions enables targeted skill development whilst providing workers with concrete career progression pathways. The visibility of employment outcomes enhances training motivation and completion rates.
Ros and Loeung (2025) analyse broader patterns in AI impact on job roles and skills requirements, confirming that effective reskilling programmes bridge skill gaps between current worker competencies and emerging position requirements. This bridging function requires close collaboration between training providers and employers to ensure curriculum relevance and graduate employability. Programmes lacking such integration may develop skills that do not align with available opportunities, limiting protective value despite educational quality.
The vacancy integration principle extends to emerging AI-specific roles, including AI operations, data analytics, robotics operation, and digital services positions. Shengelia (2025) documents how training programmes targeting these specific occupational categories demonstrate stronger employment outcomes than generic digital skills training. The specificity enables more focused skill development whilst signalling clear market demand to prospective trainees.
### Delivery models and institutional arrangements
Comparative analysis of reskilling programme delivery models reveals significant effectiveness differentials associated with institutional arrangements and governance structures. Employer-led programmes, public-private partnerships, and tripartite collaborations involving government, education, and industry consistently outperform stand-alone educational offerings in displacement prevention outcomes.
Morandini et al. (2023) comprehensively review AI impacts on worker skills, identifying corporate academies and employer-embedded training as particularly effective delivery mechanisms. These arrangements ensure curriculum relevance through direct employer input whilst providing structured pathways to internal employment. The embedded nature of training maintains worker connections to organisational systems and networks that facilitate post-training placement.
Tripartite models incorporating government policy support demonstrate enhanced outcomes across multiple national contexts. Maria et al. (2025) document how countries investing in company-based reskilling combined with AI governance frameworks experience lower AI-related unemployment than those relying solely on market-based training provision. Government involvement enables resource mobilisation beyond individual employer capacity, quality assurance mechanisms, and policy coherence across related domains including employment protection and income support.
Harsha (2025) examines public-private training partnerships and apprenticeship-style AI programmes, finding these hybrid arrangements combine employer engagement benefits with public accountability and accessibility. The partnership structure aligns incentives across stakeholders whilst enabling resource pooling and risk sharing that would challenge individual organisations.
Oladele, Orelaja and Hameed (2024) analyse workforce resilience creation in the United States context, emphasising the role of policy frameworks in supporting effective reskilling infrastructure. Their findings indicate that government investment in training capacity, combined with regulatory frameworks encouraging employer participation, creates conditions for more protective reskilling ecosystems.
Conversely, stand-alone generic online courses demonstrate limited protective capacity against AI-related redundancy. Kanagarla (2024) finds these offerings primarily benefit already-advantaged workers possessing prior digital competencies and self-directed learning capabilities. For workers most vulnerable to displacement—those with limited educational backgrounds and restricted access to learning resources—generic online training provides insufficient support to prevent redundancy. Kong, Li and Song (2025) and Zhuang (2025) reinforce these findings, documenting the adaptation gap between training provision and actual worker transition outcomes.
### Skills combinations for AI complementarity
Effective reskilling for AI-transformed workplaces requires particular combinations of digital competencies and human-centric skills that enable productive collaboration with intelligent systems. Research consistently identifies specific skill bundles demonstrating greatest protective value against displacement.
Jaiswal, Arun and Varma (2021) examine upskilling for AI in multinational corporations, identifying critical competencies including data literacy, digital tool proficiency, complex cognition, and continuous learning capacity. Their research emphasises that AI complementarity requires workers to develop capabilities that augment rather than compete with machine capabilities. This complementarity principle guides curriculum design toward skills that enhance human-AI system performance beyond what either component achieves independently.
Morandini et al. (2023) extend this analysis to broader organisational contexts, finding that effective reskilling programmes develop integrated skill profiles rather than isolated technical competencies. Their framework identifies decision-making capabilities, analytical reasoning, and adaptive problem-solving as core cognitive elements that maintain value as AI systems assume routine information processing functions.
Zirar, Ali and Islam (2023) examine human-AI coexistence in workplace settings, documenting the emergence of new skill requirements related to AI system oversight, exception handling, and collaborative task allocation. Workers in AI-augmented roles require sufficient technical understanding to interact effectively with intelligent systems whilst maintaining distinctively human judgment capabilities for complex situations.
The importance of soft skills in AI-complementary work receives substantial research attention. Depoo et al. (2025) prioritise interpersonal communication, emotional intelligence, and teamwork capabilities alongside technical competencies in their competency framework for AI-transformed roles. These human-centric skills become more rather than less important as technical tasks are automated, as workers increasingly focus on functions requiring social interaction, creative synthesis, and ethical judgment.
Cramarenco, Burcă-Voicu and Dabija (2023) systematically review AI impacts on employee skills and well-being, confirming that effective training programmes address both technical and interpersonal dimensions. Their analysis indicates that worker well-being and employment sustainability correlate with development of integrated skill profiles enabling meaningful contribution in human-AI collaborative work environments.
Li (2022) examines workforce preparation for Industry 4.0 and beyond, emphasising continuous learning capability as a meta-skill that enables ongoing adaptation as technological environments evolve. Given the pace of AI development, static skill sets face obsolescence risk, making learning agility and adaptability fundamental to sustained employability.
### Support infrastructure and programme accessibility
The design and accessibility of reskilling programmes significantly influence their protective capacity across diverse worker populations. Research documents a substantial adaptation gap where programme misalignment or inaccessibility prevents successful workforce transition despite available training opportunities.
Kanagarla (2024) quantifies this gap at approximately 42%, representing workers who experience displacement despite training availability due to barriers preventing effective participation or transition. This substantial proportion indicates that programme existence is insufficient; accessibility, appropriateness, and support infrastructure determine whether training translates into employment protection.
Kong, Li and Song (2025) analyse barriers to successful reskilling participation, identifying financial constraints, time limitations, geographic access issues, and educational prerequisites as significant obstacles. Workers most vulnerable to AI displacement often face compounded barriers, as limited educational backgrounds restrict access to training programmes whilst financial precarity constrains capacity for unpaid learning time.
Zhuang (2025) examines AI influence on labour markets with attention to worker transition mechanisms, finding that successful programmes incorporate guidance services helping workers identify appropriate training pathways and navigate complex programme landscapes. The proliferation of training options can paradoxically impede transition if workers lack information and support to select relevant opportunities.
Morandini et al. (2023) emphasise the importance of addressing age-related barriers in reskilling programme design. Older workers facing displacement often encounter particular challenges in training environments designed for younger learners, requiring pedagogical adaptation and support mechanisms addressing age-specific learning needs.
Chhibber, Rajkumar and Dassanayake (2025) analyse programme accessibility across demographic categories, finding that effective interventions provide wraparound support services including childcare, income maintenance during training, and transport assistance. These supports remove practical barriers enabling full participation, particularly for workers with caregiving responsibilities or limited financial reserves.
Cramarenco, Burcă-Voicu and Dabija (2023) document the psychological dimensions of reskilling participation, noting that anxiety, low self-efficacy, and fear of failure impede engagement, particularly among workers whose previous educational experiences were negative. Effective programmes incorporate confidence-building elements and supportive learning environments that address these psychological barriers.
Li (2022) synthesises accessibility considerations into a comprehensive framework for inclusive programme design, emphasising affordability, flexibility in scheduling and delivery mode, and recognition of prior learning as key features enabling broad participation. This framework informs both programme development and policy evaluation regarding reskilling investment effectiveness.
Discussion
The synthesis of evidence across multiple research streams enables critical evaluation of reskilling programme characteristics and their relationship to AI-related redundancy prevention. This discussion examines how the findings address each research objective whilst considering implications for theory, practice, and policy.
### Proactive versus reactive intervention
The substantial retention differential between proactive and reactive reskilling approaches—64% higher retention for proactive programmes—demands explanation and consideration of implementation implications. This finding aligns with broader human capital theory suggesting that skill development investments yield greatest returns when integrated with strategic workforce planning rather than implemented as crisis response.
The proactive advantage operates through multiple mechanisms that compound to produce substantial outcome differentials. Learning effectiveness benefits from psychological security and reduced anxiety associated with training whilst still employed. Curriculum quality improves when development timelines permit systematic needs assessment and content refinement. Employment pathway clarity enhances when training aligns with strategic workforce planning identifying future role requirements.
However, implementing proactive reskilling presents significant challenges that limit current adoption. Identifying roles facing future displacement requires sophisticated labour market intelligence and technological forecasting capabilities that many organisations lack. Investment in reskilling for roles not yet obsolete may face scrutiny regarding immediate returns, particularly in organisations under financial pressure. Furthermore, worker acceptance of reskilling for future displacement may be lower than for immediate threats, as the salience of distant risks differs from proximate dangers.
These implementation challenges suggest that policy frameworks supporting proactive reskilling—through forecasting assistance, investment incentives, and regulatory encouragement—may be necessary to realise the protective potential this approach offers. Without such support, market incentives may continue favouring reactive approaches despite their documented limitations.
### Employment linkage and programme specificity
The consistent finding that vacancy-linked programmes outperform detached training offerings carries significant implications for programme design and funding allocation. Generic skill development, however educationally valuable, demonstrates limited capacity to prevent AI-related redundancy when disconnected from actual employment opportunities.
This finding challenges assumptions underlying some public training investments, which emphasise skill development as inherently valuable regardless of employment connection. Whilst broader educational benefits may justify such investments on alternative grounds, displacement prevention specifically requires tight integration with employer requirements and placement pathways.
The vacancy integration principle suggests that effective reskilling ecosystems require infrastructure connecting training providers with employers, ensuring curriculum relevance and graduate placement. Such infrastructure may include labour market information systems, employer advisory arrangements, and placement services that bridge training completion and employment commencement.
For workers in organisations implementing hiring freezes, the employment linkage finding suggests that internal reskilling pipelines may offer greater protection than external training programmes, as internal vacancies may continue despite external hiring restrictions. Organisations maintaining internal mobility whilst restricting external recruitment create opportunities for protected worker transitions that external programmes cannot provide.
### Institutional arrangements and governance
The superior outcomes associated with employer-led, public-private partnership, and tripartite delivery models relative to stand-alone educational offerings reflect the importance of institutional arrangements in reskilling effectiveness. These findings suggest that governance structures—the relationships between training providers, employers, government, and workers—significantly influence programme outcomes beyond curriculum content or pedagogical quality.
Tripartite arrangements enable resource mobilisation and risk sharing that would challenge individual actors. Government can provide financial resources, quality assurance, and policy coherence. Employers contribute labour market intelligence, placement opportunities, and training relevance. Educational institutions offer pedagogical expertise and credentialing capacity. Workers and their representatives ensure programme accessibility and responsiveness to learner needs.
The weakness of stand-alone online courses in preventing redundancy challenges the substantial investment directed toward such offerings by both private providers and public programmes. Whilst these courses may serve supplementary functions for already-advantaged learners, their limited protective capacity for vulnerable workers suggests misallocation of resources intended for displacement prevention.
This finding does not preclude value from online learning modalities, but rather emphasises that delivery format matters less than institutional arrangements connecting training to employment. Online programmes integrated within employer-led or partnership frameworks may be effective; the same content offered without such integration demonstrates limited impact.
### Skill combinations for contemporary workforce protection
The emphasis on integrated skill profiles combining digital competencies with soft skills reflects the changing nature of human contribution in AI-augmented work environments. As AI systems assume routine cognitive tasks, human workers add value through capabilities that remain distinctively human—interpersonal communication, ethical judgment, creative synthesis, and emotional intelligence.
This skill combination requirement challenges traditional vocational training models that emphasise technical competency development as primary. Effective preparation for AI-complementary roles requires attention to human-centric capabilities that may receive limited emphasis in technically oriented curricula. Reskilling programme design must balance technical skill development with attention to interpersonal and cognitive capabilities that maintain value as automation advances.
The continuous learning meta-skill identified in the literature carries particular implications for programme design. Single-episode training preparing workers for specific roles may provide temporary protection but leaves workers vulnerable to subsequent technological change. Developing capacity for ongoing skill acquisition enables sustained adaptation as work environments continue evolving, providing more durable employment protection.
### Accessibility and the adaptation gap
The documented 42% adaptation gap—workers who fail to transition successfully despite available training—represents a significant challenge to reskilling as a displacement prevention strategy. This substantial proportion suggests that programme availability is insufficient; effectiveness requires addressing barriers preventing successful participation and completion.
The composition of this adaptation gap likely concentrates workers already facing labour market disadvantage—those with limited educational backgrounds, financial constraints, caregiving responsibilities, and psychological barriers to learning engagement. Ironically, these workers may face greatest AI displacement risk whilst encountering the most significant barriers to protective training access.
Closing this gap requires programme design explicitly addressing accessibility across multiple dimensions. Financial support during training maintains income for workers who cannot afford unpaid learning time. Childcare and transport assistance removes practical barriers constraining participation. Recognition of prior learning enables accelerated progression for workers with relevant experiential knowledge. Supportive learning environments address anxiety and low self-efficacy that impede engagement.
The accessibility imperative suggests that cost-benefit analysis of reskilling investments must consider not merely training delivery costs but also support service requirements to enable effective participation. Programmes achieving high completion rates may require substantially greater per-participant investment than minimum-service offerings, but this investment yields actual displacement prevention rather than merely documented training provision.
### Framework for protective programme classification
Synthesising across these dimensions, a framework emerges for classifying reskilling programmes according to protective capacity. Strongly protective programmes demonstrate multiple characteristics: proactive timing, tight linkage to specific vacancies and redesigned roles, employer or tripartite governance arrangements, development of integrated technical and soft skill profiles, and accessible design with comprehensive support services.
Moderately protective programmes may exhibit some but not all of these characteristics—for instance, employer-linked training that remains reactive rather than proactive, or proactive programmes that lack adequate accessibility provisions. These programmes offer meaningful but limited protection, with effectiveness varying according to worker circumstances and organisational contexts.
Weakly protective programmes—exemplified by generic stand-alone online courses without employment linkage or support services—provide limited displacement prevention despite potentially high educational quality. These offerings may benefit already-advantaged workers capable of navigating learning and labour market systems independently, but offer inadequate support for workers most vulnerable to AI-related redundancy.
This classification framework enables more precise evaluation of reskilling investments and more targeted resource allocation toward genuinely protective interventions. Rather than treating all training as equivalent, the framework distinguishes programme types according to demonstrated effectiveness in preventing displacement outcomes.
Conclusions
This dissertation has systematically examined the characteristics of reskilling programmes that genuinely protect workers from AI-related redundancy, addressing a critical gap in understanding as organisations navigate technological transformation whilst implementing restrictive hiring practices. The analysis establishes clear conclusions regarding each research objective.
Regarding timing of intervention, the evidence conclusively demonstrates that proactive reskilling substantially outperforms reactive training, with organisations initiating programmes before displacement imminence achieving retention rates 64% higher than those responding only to immediate redundancy threats. This finding underscores the importance of anticipatory workforce planning and early intervention in protection strategies.
The analysis confirms that tight linkage to real vacancies and redesigned roles critically determines programme effectiveness. Training that pipelines workers into specific emerging positions—AI operations, data analytics, digital services, and robotics operation—achieves substantially better displacement prevention than generic skill development disconnected from employment opportunities.
Comparative evaluation of delivery models reveals the superior effectiveness of employer-led initiatives, public-private partnerships, and tripartite arrangements incorporating government, education, and industry stakeholders. Stand-alone generic training, particularly online courses without institutional support, demonstrates limited protective capacity despite potentially serving supplementary educational functions.
The skill combinations providing greatest protection integrate digital competencies with soft skills—communication, interpersonal capability, emotional intelligence, and teamwork—that remain distinctively human as AI systems automate routine cognitive tasks. Effective programmes develop integrated skill profiles and, crucially, continuous learning capacity enabling ongoing adaptation.
Finally, the analysis documents a substantial 42% adaptation gap where programme inaccessibility or misalignment prevents successful workforce transition. Closing this gap requires comprehensive support infrastructure addressing financial, practical, and psychological barriers, with particular attention to workers facing compounded disadvantages in learning engagement.
The significance of these findings extends across academic, policy, and practice domains. Theoretically, this research contributes to understanding of human capital formation under technological change, emphasising the institutional and contextual factors that shape skill development effectiveness beyond individual learning outcomes. For policy, the findings inform more discriminating evaluation of public reskilling investments, directing resources toward genuinely protective programme types rather than treating all training as equivalent. For organisational practice, the framework developed guides strategic workforce development decisions, enabling targeted investment in approaches demonstrated to prevent redundancy rather than merely documenting training activity.
Future research should extend this analysis in several directions. Longitudinal studies tracking worker outcomes over extended periods would establish durability of protective effects and identify factors influencing sustained employment following reskilling completion. Comparative analysis across national policy contexts would illuminate the role of regulatory frameworks and public investment patterns in shaping reskilling ecosystem effectiveness. Investigation of psychological and motivational factors influencing reskilling engagement would inform more effective programme design addressing participation barriers. Finally, as AI technologies continue advancing, ongoing research must track evolving skill requirements and protective programme characteristics, ensuring that guidance remains current with technological conditions.
In conclusion, this dissertation establishes that reskilling programmes most reliably protecting workers from AI-related redundancy are proactive, employer-connected initiatives that pipeline workers into clearly defined AI-complementary roles, combine digital and soft skill development, and are backed by policy support and accessible design. Generic, late-stage training offers limited protection regardless of educational quality, underscoring the importance of institutional arrangements, timing, and employment linkage in determining programme effectiveness. As organisations navigate AI transformation during periods of hiring constraint, these findings provide evidence-based guidance for intervention strategies that genuinely protect workers from technological displacement.
References
Chhibber, S., Rajkumar, S. and Dassanayake, S., 2025. Will Artificial Intelligence Reshape the Global Workforce by 2030? A Cross-Sectoral Analysis of Job Displacement and Transformation. *Blockchain, Artificial Intelligence, and Future Research*. https://doi.org/10.70211/bafr.v1i1.178
Costa, I., Machado, C. and Conceição, O., 2024. Automation and the Importance of Reskilling Workers: A Case Study in the Automotive Industry. *European Conference on Management Leadership and Governance*. https://doi.org/10.34190/ecmlg.20.1.2923
Cramarenco, R., Burcă-Voicu, M. and Dabija, D., 2023. The impact of artificial intelligence (AI) on employees’ skills and well-being in global labor markets: A systematic review. *Oeconomia Copernicana*. https://doi.org/10.24136/oc.2023.022
Depoo, L., Hajerová-Mullerová, L., Kronberger, Z., Říhová, G., Stříteský, M., Hořáková, M., Legnerová, K., Palíšková, M., Němec, O., Šmíd, D., Jurčík, T. and Kopecký, M., 2025. The Impact of AI Implementation on Job Transformation and Competency Requirements: Prioritising Reskilling and Soft Skills Development. *Quality Innovation Prosperity*. https://doi.org/10.12776/qip.v29i2.2165
Harsha, S., 2025. Navigating the Future of Work: The Impact of Artificial Intelligence on Jobs, Skills, and Workforce Dynamics. *International Journal of Business and Management Invention*. https://doi.org/10.35629/8028-14056779
Jaiswal, A., Arun, C. and Varma, A., 2021. Rebooting employees: upskilling for artificial intelligence in multinational corporations. *The International Journal of Human Resource Management*, 33, pp. 1179-1208. https://doi.org/10.1080/09585192.2021.1891114
Kanagarla, K., 2024. Artificial Intelligence and Employment Transformation: A Multi-Sector Analysis of Workforce Disruption and Adaptation. *International Journal of Scientific Research in Computer Science, Engineering and Information Technology*. https://doi.org/10.32628/cseit24106170
Kong, Z., Li, Z. and Song, Y., 2025. Analysis of Artificial Intelligence-Driven Job Replacement in the Service Industry and Unemployment Response Strategies. *Advances in Economics, Management and Political Sciences*. https://doi.org/10.54254/2754-1169/2025.bl23302
Li, L., 2022. Reskilling and Upskilling the Future-ready Workforce for Industry 4.0 and Beyond. *Information Systems Frontiers*, pp. 1-16. https://doi.org/10.1007/s10796-022-10308-y
Maria, S., Purwinahyu, P., Fitriansyah, F., Rachmawaty, A. and Aini, R., 2025. Artificial Intelligence and Labor Markets: Analyzing Job Displacement and Creation. *International Journal of Engineering, Science and Information Technology*. https://doi.org/10.52088/ijesty.v5i2.830
Morandini, S., Fraboni, F., Angelis, M., Puzzo, G., Giusino, D. and Pietrantoni, L., 2023. The Impact of Artificial Intelligence on Workers’ Skills: Upskilling and Reskilling in Organisations. *Informing Science: The International Journal of an Emerging Transdiscipline*, 26, pp. 39-68. https://doi.org/10.28945/5078
Oladele, I., Orelaja, A. and Hameed, A., 2024. Artificial intelligence and automation: creating a more resilient United States workforce. *Journal of Scientific Papers “Social Development and Security”*. https://doi.org/10.33445/sds.2024.14.4.7
Organisation for Economic Co-operation and Development, 2023. *OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market*. Paris: OECD Publishing.
Ros, P. and Loeung, B., 2025. Analyzing the Impact of Artificial Intelligence on Job Roles, Skills Requirements, and Employment Patterns. *Journal of Social Sciences and Humanities*. https://doi.org/10.56943/jssh.v4i1.699
Shengelia, R., 2025. Artificial Intelligence and Labor Market Dynamics: Employment Problems and Development Trends. *Economics*. https://doi.org/10.36962/ecs107/3-5/2025-07
World Economic Forum, 2023. *Future of Jobs Report 2023*. Geneva: World Economic Forum.
Zhuang, Y., 2025. The Influence of Artificial Intelligence on Labor Markets. *SHS Web of Conferences*. https://doi.org/10.1051/shsconf/202521803030
Zirar, A., Ali, S. and Islam, N., 2023. Worker and workplace Artificial Intelligence (AI) coexistence: Emerging themes and research agenda. *Technovation*. https://doi.org/10.1016/j.technovation.2023.102747
