How AI can Transform HR and People Analytics: Privacy, Transparency and Trust issues
Barry Swales · 29/03/2023
Artificial intelligence (AI) is a powerful tool that can help HR professionals and people analytics teams to improve their processes, insights and outcomes. AI can automate tasks, enhance decision making, and provide personalized recommendations for employees and managers. However, AI also poses some challenges and risks that need to be addressed, especially when it comes to privacy, transparency and trust.
Privacy: How can HR protect personal data in an AI-driven world?
One of the main concerns about AI is how it collects, uses and shares personal data of employees and candidates. Personal data is any information that can identify or relate to a person, such as name, email, phone number, biometric data, health records, performance reviews, etc. Personal data is valuable for AI because it can enable more accurate and relevant analysis and predictions. However, personal data is also sensitive and subject to legal and ethical regulations and expectations.
HR professionals and people analytics teams need to ensure that they respect the privacy rights and preferences of their data subjects, and comply with the applicable laws and standards in their jurisdictions. For example, in the European Union, the General Data Protection Regulation (GDPR) sets strict rules for how personal data can be processed and transferred by organizations. In the United States, different states have different laws regarding data privacy, such as the California Consumer Privacy Act (CCPA).
Some of the best practices for privacy protection in AI include:
- Obtaining informed consent from data subjects before collecting and using their personal data for AI purposes
- Minimising the amount and sensitivity of personal data collected and used for AI purposes
- Anonymising or pseudonymising personal data whenever possible to reduce the risk of re-identification
- Encrypting or hashing personal data when storing or transmitting it to prevent unauthorised access or disclosure
- Implementing data retention policies and deleting personal data when it is no longer needed or requested by data subjects
- Conducting privacy impact assessments and audits to identify and mitigate potential privacy risks and breaches
- Providing data subjects with access, correction, deletion and portability rights over their personal data
- Establishing clear roles and responsibilities for data protection within the organisation and with external partners.
Transparency: How can HR explain the logic and outcomes of AI?
Another concern about AI is how it operates and produces results that affect employees and candidates. AI models are often complex and opaque, making it difficult to understand how they work and why they make certain decisions or recommendations. This lack of transparency can lead to confusion, frustration, mistrust or even discrimination among data subjects.
Another concern about AI is how it operates and produces results that affect employees and candidates. AI models are often complex and opaque, making it difficult to understand how they work and why they make certain decisions or recommendations. This lack of transparency can lead to confusion, frustration, mistrust or even discrimination among data subjects.
HR professionals and people analytics teams need to ensure that they provide clear and meaningful explanations of their AI models and outcomes to their data subjects and stakeholders. Transparency can help increase trust, accountability and fairness in AI. Transparency can also help identify and correct errors or biases in AI models or outcomes.
Some of the best practices for transparency in AI include:
- Documenting the purpose, scope, methods, assumptions, limitations and sources of data of AI models
- Providing simple and intuitive visualisations or summaries of AI models and outcomes
- Using techniques such as feature importance, sensitivity analysis or counterfactuals to highlight the key factors or variables that influence AI outcomes
- Using techniques such as local interpretable model-agnostic explanations (LIME) or SHapley Additive exPlanations (SHAP) to generate local or global explanations of AI outcomes
- Providing feedback mechanisms for data subjects to ask questions, express concerns or challenge AI outcomes
- Testing and validating AI models and outcomes against predefined criteria or benchmarks
- Disclosing any uncertainties, risks or trade-offs associated with AI models or outcomes.
Trust: How can HR build confidence and acceptance of AI?
A third concern about AI is how it affects the relationship between employees, candidates, managers and HR professionals. AI can create new opportunities for collaboration, communication and empowerment among these actors. However, AI can also create new challenges for trust, engagement and satisfaction among these actors.
HR professionals and people analytics teams need to ensure that they foster a positive and supportive culture around AI within their organization. Trust is essential for the successful adoption and use of AI in HR. Trust can help overcome resistance, fear or skepticism among employees or candidates towards AI. Trust can also help enhance loyalty, motivation or performance among employees or managers towards HR.
Some of the best practices for trust building in AI include:
HR professionals need to ensure that they design AI systems that are human-centric, empathetic and respectful of employees' preferences, emotions and values. They also need to ensure that employees can trust the reliability, validity and quality of the information and advice provided by AI systems, and that they can verify their sources and accuracy. Moreover, they need to be transparent about the goals and intentions of using AI for employee engagement, and how they protect employee autonomy, dignity and privacy.
AI offers tremendous opportunities for HR professionals to improve workforce analytics and HR digitalisation. However, it also raises important issues and challenges for privacy, transparency and trust that need to be addressed carefully and ethically. HR professionals need to adopt a responsible approach to AI implementation that respects employee rights, expectations and interests, and that fosters a culture of openness, accountability and collaboration.
Register now for the annual global conference for People Analytics and Data-Driven HR
19-20 April · QEII Centre · London