AI refers to computer systems that perform tasks normally associated with human intelligence, such as recognising patterns, generating text, predicting outcomes or making recommendations. In people practice, AI may be used to summarise policies, draft job adverts, identify skills gaps, match candidates to roles, support onboarding, analyse absence trends, personalise learning or answer employee queries. Generative AI can create new content, while predictive tools use data to estimate future outcomes. Neither should be treated as neutral or automatically correct.
The opportunity for the people profession is significant. AI can reduce repetitive administration, improve access to information and help practitioners focus on higher-value work such as diagnosis, coaching, workforce planning and organisational change. People analytics is already linked to evidence-based decision-making in the CIPD Profession Map, where practitioners are expected to use people data responsibly to create insight and value (CIPD, 2024). AI can strengthen this work by helping teams explore larger data sets, spot patterns and test assumptions. For example, an AI-enabled analytics tool might highlight turnover hotspots, but a people professional still needs to interpret the causes by considering workload, management practice, pay, culture, labour market conditions and employee voice.
Responsible use starts with purpose. A weak process does not become fair or effective because AI is added to it. If a recruitment process relies on unclear criteria, inconsistent scoring or biased historical data, an AI tool may reproduce the same weaknesses at greater scale. Research on AI-enabled recruitment highlights both benefits, such as efficiency and consistency, and risks, including discrimination linked to gender, race, personality traits or other protected characteristics (Chen, 2023). The EHRC (2024) also warns that AI may create equality and human rights risks, so fairness cannot be treated as a technical issue only. This is why the UK Government’s responsible AI in recruitment guidance stresses assurance, transparency, fairness, accountability, contestability and redress when procuring or deploying AI systems (DSIT, 2024).
Data protection is also central. People data is sensitive because it can reveal personal, behavioural, performance, health, equality or employment information. The ICO (2023) explains that AI systems must still meet data protection principles, including fairness, transparency, data minimisation, accuracy, security and accountability. Under UK GDPR, solely automated decisions that have legal or similarly significant effects on individuals are restricted unless specific conditions and safeguards apply (ICO, nd). In people practice, this is especially relevant to recruitment, promotion, performance management, disciplinary processes, redundancy selection and access to development opportunities.
Human oversight must be meaningful rather than symbolic. A manager or HR adviser should understand what the system has done, what data it used, where uncertainty exists and when to challenge the output. Kellogg, Valentine and Christin (2020) show that algorithmic systems can reshape control at work by directing, evaluating and disciplining workers through data. For people professionals, this creates a responsibility to question whether AI improves work or simply intensifies monitoring, pressure and mistrust. Acas (2025) advises employers to consult staff and representatives, develop clear AI policies, check outputs for accuracy, tone and bias, and protect sensitive or personal information.
Good governance turns ethical principles into routine practice. NIST (2023) frames AI risk management around mapping, measuring, managing and governing risk. ISO/IEC 42001:2023 provides an international management system standard for organisations developing or using AI, with emphasis on responsible use, transparency, reliability and continual improvement (ISO, 2023). In practical terms, people professionals should require clear ownership, risk assessment, equality impact consideration, data protection impact assessment where needed, supplier assurance, user training, audit trails and review dates.
AI should therefore be viewed as a decision-support tool, not a substitute for professional judgement. The people professional’s role is to ask better questions: Is the use case legitimate? Is the data suitable? Could the tool disadvantage any group? Can the decision be explained? Can employees challenge it? Is there a trained human accountable for the final judgement? When these questions are built into digital working, AI can support more responsive, evidence-informed and inclusive people practice.
Action Point
Select one people process where AI is being used, considered or informally tested. Map the purpose, data used, people affected, likely benefits, possible harms and decision points requiring human judgement. Then identify one improvement that would make the process more transparent, fair, secure or evidence-based.