A new study has highlighted the potential of data analytics and artificial intelligence to identify structural inefficiencies and hidden biases in workforce development systems, recommending their integration into policymaking to promote fairness, improve employment opportunities and strengthen economic growth. The study, titled “Detecting Structural Inefficiencies and Implicit Bias in Workforce……
A new study has highlighted the potential of data analytics and artificial intelligence to identify structural inefficiencies and hidden biases in workforce development systems, recommending their integration into policymaking to promote fairness, improve employment opportunities and strengthen economic growth.
The study, titled “Detecting Structural Inefficiencies and Implicit Bias in Workforce Development Systems: A Data Analytics Framework for Equitable Policy Design,” was authored by Goodness Rex Nze-Igwe and Chinyere Agbasiere.
Published in the July 2026 edition of Path of Science, the conceptual paper examines how governments, employers and workforce development institutions can use data analytics to uncover systemic barriers that prevent individuals, particularly marginalised groups, from accessing training, employment and career advancement opportunities.
The researchers argue that although workforce development programmes are designed to reduce unemployment, promote economic mobility and strengthen national productivity, many continue to suffer from fragmented data systems, outdated training programmes, bureaucratic bottlenecks and implicit biases.
According to the study, these challenges frequently remain undetected because traditional evaluation methods focus on aggregate performance indicators, such as the number of people trained or placed in employment, rather than examining which groups benefit from the programmes and which remain excluded.
The paper identifies women, ethnic minorities, persons with disabilities, rural residents, informal-sector workers and other disadvantaged populations among those who may experience barriers to participation in workforce development initiatives.
It notes that seemingly neutral policies, including rigid educational requirements, inflexible training schedules and restrictive eligibility criteria, can unintentionally exclude people with relevant skills or those unable to meet conventional programme requirements.
The researchers also warn that the concentration of training centres and employment support services in urban areas can disadvantage people in rural and remote communities, reinforcing existing social and economic inequalities.
Four analytical approaches proposed
To address these challenges, the study proposes a conceptual framework that integrates four forms of data analytics: descriptive, diagnostic, predictive and prescriptive analytics.
Descriptive analytics examines historical data to identify trends in employment rates, training completion, programme participation and workforce demographics. This can help institutions understand existing patterns and assess programme performance.
Diagnostic analytics goes further by investigating the underlying causes of observed disparities. For instance, it can help explain why women participate less in technical training programmes or why certain groups experience higher dropout rates.
Predictive analytics uses historical information, statistical modelling and machine learning to anticipate future workforce trends, skills shortages and the likelihood of programme participants dropping out or experiencing prolonged unemployment.
Prescriptive analytics, meanwhile, uses these insights to recommend practical interventions, including improved training programmes, more effective resource allocation and tailored career pathways.
The authors explain that combining these approaches can help policymakers move beyond reacting to existing problems towards anticipating challenges and designing interventions that address their underlying causes.
AI and workforce policy
The study highlights the growing role of artificial intelligence, machine learning, data visualisation and labour market information systems in improving workforce planning and decision-making.
It observes that machine learning can identify patterns in employment histories, training participation and demographic information, while natural language processing and sentiment analysis can help institutions examine employee feedback, exit interviews and programme evaluations.
Labour Market Information Systems can also help policymakers track changing employment demands, identify skills gaps and align training programmes with employers’ needs.
However, the researchers caution that technology alone cannot guarantee fairness.
They argue that artificial intelligence systems trained on biased historical data may reproduce existing inequalities, potentially favouring groups that have traditionally enjoyed greater access to employment while disadvantaging underrepresented populations.
For example, recruitment algorithms trained on historical hiring records may replicate previous preferences for particular demographic groups. Similarly, training eligibility systems that rely heavily on formal educational qualifications may overlook individuals who possess practical skills but lack conventional credentials.
The study therefore recommends that organisations regularly audit algorithms, make automated decisions more transparent and ensure that people can question decisions that affect their access to training and employment.
Three theories underpin the framework
The researchers ground their proposed framework in Data Justice Theory, Institutional Theory and Intersectionality.
Data Justice Theory emphasises fairness in how information is collected, analysed and used, particularly where decisions affect vulnerable communities. It also stresses privacy, informed consent, transparency and the participation of affected populations in data governance.
Institutional Theory explains how established organisational practices, cultural norms and power relationships can sustain inefficient or exclusionary policies, even when new technologies are introduced.
Intersectionality, meanwhile, examines how overlapping factors such as gender, socioeconomic status, ethnicity, age and disability can combine to create different experiences of disadvantage.
The authors argue that policymakers should examine these factors together rather than treating population groups as uniform categories. Such an approach, they say, can reveal inequalities that broad performance indicators might conceal.
Recommendations for policymakers and employers
The paper calls for greater investment in modern data infrastructure, staff training, cross-sector collaboration and more comprehensive labour market information systems.
It recommends stronger cooperation among government agencies, employers, training institutions and civil society organisations to ensure that workforce programmes reflect changing labour market needs and the experiences of the people they serve.
The study also advocates the use of disaggregated data to identify differences in access to training, programme completion, employment outcomes and career progression across demographic groups.
In addition, it urges policymakers to revise outdated eligibility requirements, improve training curricula, redistribute resources to underserved communities and establish continuous monitoring mechanisms to evaluate the effectiveness of interventions.
The researchers emphasise that affected communities, including job seekers and training participants, should be involved in designing and evaluating data systems. Such participation, they argue, can improve public trust, make policies more relevant and reveal barriers that administrative records alone may fail to capture.
The paper nevertheless acknowledges the risks associated with expanding data-driven workforce management, including excessive workplace surveillance, privacy violations, discriminatory profiling and the exclusion of communities that lack adequate digital infrastructure.
It therefore calls for robust privacy protections, ethical data practices, algorithmic accountability and targeted investment to prevent the digital divide from widening existing inequalities.
Call for further research
The researchers conclude that data analytics can strengthen workforce development systems when combined with ethical governance, institutional reform and a commitment to social inclusion.
They maintain that the proposed framework offers a pathway for identifying hidden inefficiencies, improving the allocation of resources and designing policies that respond more effectively to labour market needs.
However, the paper is conceptual and proposes a framework rather than reporting the results of an empirical field test. The authors recommend that future research apply the model to specific workforce programmes and policy settings to determine whether data-driven interventions can measurably reduce inequality and improve employment outcomes.
