Diverse teams can improve the range of perspectives involved in building artificial intelligence, but representation alone does not guarantee that an AI system will be fair, accurate or ethical.<\/p>\n
Responsible AI requires both inclusive participation<\/strong> and risk-management discipline<\/strong>: clear objectives, appropriate data, testing across relevant groups, transparency, accessibility, human oversight, impact assessment and monitoring after deployment.<\/p>\n This is especially important when AI is used in employment, where automated systems can influence who sees a vacancy, who is shortlisted, how candidates are assessed, how workers are scheduled and how performance is evaluated.<\/p>\n AI systems are socio-technical systems. They are shaped not only by code, but by decisions about:<\/p>\n A team with a narrow range of experience may be more likely to miss an important user group, accessibility issue, cultural assumption or harmful edge case.<\/p>\n UNESCO’s Recommendation on the Ethics of Artificial Intelligence places diversity, inclusion, fairness, transparency and human oversight among the core principles of responsible AI.<\/p>\n Source: UNESCO, Recommendation on the Ethics of Artificial Intelligence<\/a><\/p>\n It is possible to have a demographically diverse team and still build a harmful system. It is also possible for a relatively homogeneous team to identify some risks through strong governance and external challenge.<\/p>\n The stronger approach is to combine representation with:<\/p>\n Before looking at the model, ask what the system has been told to optimise.<\/p>\n The International Labour Organization’s 2025 working paper on AI in human-resource management warns that AI systems can encode poorly defined objectives, rely on biased or incomplete data and operate through opaque processes.<\/p>\n Source: ILO, AI in Human Resource Management: The Limits of Empiricism, 2025<\/a><\/p>\n A recruitment system designed to predict “culture fit”, “growth mindset” or “future potential” may appear objective while operationalising a vague human concept through weak proxies.<\/p>\n Ask:<\/p>\n AI systems learn patterns from data. If historical data reflects unequal access, biased decisions or underrepresentation, the system can reproduce those patterns.<\/p>\n Data problems can include:<\/p>\n The question is not only whether the dataset is large. It is whether the data is appropriate for the decision the system is being asked to support.<\/p>\n Hiring and workforce-management systems affect access to jobs, income and career opportunity.<\/p>\n The ILO’s 2025 analysis of AI in HR highlights risks in recruitment, compensation, scheduling and performance management, particularly when systems are built on reductive objectives or opaque data-driven assumptions.<\/p>\n The ILO also notes that AI can potentially support fairer and more transparent HR processes when it is designed and governed responsibly.<\/p>\n When recruiting AI professionals, organisations can reduce arbitrary decision-making by using:<\/p>\n Diversity goals should not replace job-related assessment. The objective is to widen opportunity and reduce unnecessary barriers while maintaining a clear standard for the work.<\/p>\n AI teams need a range of roles, not only machine-learning researchers.<\/p>\n Depending on the system, relevant expertise can include:<\/p>\n Overly narrow degree, pedigree or experience requirements can shrink the talent pool without improving job performance.<\/p>\n An engineer may identify model-performance issues. A domain expert may recognise that the target variable is meaningless in practice. An accessibility specialist may identify a barrier for disabled users. A privacy professional may identify unnecessary data collection. A frontline employee may show that the workflow does not match real work.<\/p>\n This is why NIST describes AI risk management as a multidisciplinary and socio-technical activity rather than a purely technical exercise.<\/p>\n Source: NIST AI Risk Management Framework<\/a><\/p>\n Internal diversity is useful, but the people building the system are not always the people most affected by it.<\/p>\n Depending on the use case, stakeholder participation can include:<\/p>\n Participation is most useful when stakeholders can influence design choices rather than being consulted after the important decisions have already been made.<\/p>\n An overall accuracy score can hide important differences.<\/p>\n Responsible testing may need to examine:<\/p>\n Which groups are appropriate to test depends on the use case, population and law.<\/p>\n AI-enabled recruitment can create barriers for candidates with disabilities when systems assume a narrow range of speech, facial movement, interaction style, device use or response format.<\/p>\n The ILO has highlighted that AI can both improve accessibility and create exclusion depending on how systems are designed and used.<\/p>\n Source: ILO, AI’s double-edged sword and disability employment<\/a><\/p>\n Organisations should provide accessible alternatives and a process for reasonable accommodation appropriate to the jurisdiction.<\/p>\n An automated system can reproduce human bias through historical data, labels, proxies or design choices. It can also create new forms of bias through optimisation and scale.<\/p>\n Human review is still important, but “human in the loop” is not enough if the reviewer simply accepts the system’s output.<\/p>\n Reviewers need:<\/p>\n UNESCO’s AI ethics framework emphasises that AI should not displace ultimate human responsibility and accountability.<\/p>\n Meaningful oversight asks:<\/p>\n UNESCO has developed an Ethical Impact Assessment tool to help organisations examine alignment with its AI ethics principles.<\/p>\n Source: UNESCO, Ethical Impact Assessment<\/a><\/p>\n Before deploying a consequential system, assess:<\/p>\n A system that performs acceptably in testing can behave differently when:<\/p>\n Responsible AI requires ongoing monitoring, not a one-time fairness test.<\/p>\n Buying a recruitment or HR AI product does not remove the organisation’s responsibility to understand what it is doing.<\/p>\n Ask vendors:<\/p>\n If an organisation wants a more representative AI workforce, recruitment is only the first step.<\/p>\n Track, where lawful and appropriate:<\/p>\n A company can hire a diverse entry-level cohort and still lose diversity through promotion or retention patterns.<\/p>\n Representation is less useful if people do not feel safe challenging a product decision.<\/p>\n Leaders should make it legitimate to ask:<\/p>\n For a consequential AI project, consider whether the team has access to:<\/p>\n The exact mix depends on the system and risk.<\/p>\n Diversity can make AI teams better able to see risks and serve a wider range of users. But ethical outcomes require an operating system around that diversity: governance, testing, accountability and real stakeholder participation.<\/p>\n The strongest question is therefore not “Is our AI team diverse?” It is “Do we have the perspectives, processes and controls needed to identify and manage the risks this system creates?”<\/p>\n MATSH provides professional-development content across leadership, HR, diversity, technology and workplace capability. Organisations adopting AI in HR should combine technical expertise with strong people-management, fairness and governance capability.<\/p>\n No. Diversity can broaden perspective, but fairness also depends on objectives, data, testing, governance, oversight and monitoring.<\/p>\n Bias can enter through historical data, labels, proxies, poorly defined objectives, non-representative testing or the way the system is deployed.<\/p>\n AI can support recruitment, but employers should assess whether the tool is job-related, accurate, accessible, transparent enough for the use case and governed appropriately under applicable law.<\/p>\n Humans should retain meaningful responsibility for consequential decisions, understand system limitations, be able to override outputs and provide a route for challenge or redress.<\/p>\n There is no universal structure. Consequential projects often need technical, domain, privacy, security, legal, accessibility, human-factors and affected-stakeholder perspectives.<\/p>\n Evidence-led guide to diverse AI teams and responsible AI, covering data quality, testing, accessibility, human oversight, impact assessment and ongoing monitoring.<\/p>\n","protected":false},"author":1,"featured_media":5236,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_kad_post_transparent":"","_kad_post_title":"","_kad_post_layout":"","_kad_post_sidebar_id":"","_kad_post_content_style":"","_kad_post_vertical_padding":"","_kad_post_feature":"","_kad_post_feature_position":"","_kad_post_header":false,"_kad_post_footer":false,"_kad_post_classname":"","footnotes":""},"categories":[268,269],"tags":[],"class_list":["post-5235","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-human-resources","category-recruitment"],"_links":{"self":[{"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/posts\/5235","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/comments?post=5235"}],"version-history":[{"count":5,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/posts\/5235\/revisions"}],"predecessor-version":[{"id":10312,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/posts\/5235\/revisions\/10312"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/media\/5236"}],"wp:attachment":[{"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/media?parent=5235"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/categories?post=5235"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/matsh.co\/en\/wp-json\/wp\/v2\/tags?post=5235"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}Why diversity still matters in AI development<\/h2>\n
\n
Diverse teams are a safeguard, not a guarantee<\/h2>\n
\n
Start with the objective the AI system is optimising<\/h2>\n
\n
Data quality is an ethical issue<\/h2>\n
\n
Employment AI deserves especially careful scrutiny<\/h2>\n
Inclusive recruitment teams should not rely on intuition alone<\/h2>\n
\n
Remove unnecessary barriers from AI job descriptions<\/h2>\n
\n
Multidisciplinary teams can see different classes of risk<\/h2>\n
Include affected people, not only internal experts<\/h2>\n
\n
Test performance across relevant groups<\/h2>\n
\n
Accessibility should be designed in from the start<\/h2>\n
Do not assume automation removes human bias<\/h2>\n
\n
Human oversight should be meaningful<\/h2>\n
\n
Use impact assessment before high-risk deployment<\/h2>\n
\n
Monitor the system after deployment<\/h2>\n
\n
Vendor tools still require employer accountability<\/h2>\n
\n
Diversity metrics should cover the employee lifecycle<\/h2>\n
\n
Create a team culture where risk can be raised<\/h2>\n
\n
A practical responsible-AI staffing model<\/h2>\n
\n
Responsible AI is a process, not a diversity statement<\/h2>\n
MATSH and responsible workplace technology<\/h2>\n
Frequently asked questions<\/h2>\n
Does a diverse AI team automatically produce unbiased AI?<\/h3>\n
Why can AI hiring tools be biased?<\/h3>\n
Should employers use AI in hiring?<\/h3>\n
What is the role of human oversight?<\/h3>\n
What should an AI ethics team include?<\/h3>\n
Related MATSH resources<\/h2>\n
\n
Sources<\/h2>\n
\n