An HR team uses an AI solution to assess job applicants or identify employees that might be considering leaving the organization. While the system is managed by IT, the decisions that result from its usage have implications for trust within the organization. In case the system results in discriminatory advice, then the problem no longer falls under the domain of IT, but of HR.
This article explains the contribution of HR to AI governance.
How HR Helps in AI Governance Where IT Fails
For HR AI governance, there needs to be an evaluation of the impact of technology on employees. HR appreciates the environment that hiring, performance appraisal, staffing, and employee relations operate in.
HR also brings the employee perspective into responsible AI in HR. Before an AI system is deployed, HR can assess whether employees understand how their data is being used and whether affected employees have a path to question or appeal an AI decision.
HR’s Understanding of Culture Can Be an Asset for AI Governance
Organizational culture affects employee perception of AI systems. It is important to AI governance in HR because AI can pose different risks based on how employees perceive AI decisions.
HR can utilize this information to create governance standards that reflect how AI works among other employees. This includes identifying decisions that require human review, setting expectations for manager accountability, and establishing communication around AI.
This cultural perspective also helps HR identify risks that technical monitoring may be missed. For example, the organization can develop automation bias even when the underlying model performs as expected. By incorporating employee feedback and manager input into governance reviews, HR can make it responsive to workplace conditions.
Creating the Cross-Functional Committee on AI Governance
Step 1: Identify the Scope of the Committee
First, identify the scope of duties of the committee. The mandate must include all AI applications that affect employees. It will be useful to specify if the committee can either allow, disallow, suspend, or modify any application of AI.
Committee approval is necessary before any AI solution deployed to rank candidates and recommend promotions is put into use.
Step 2: Develop an AI Risk Framework
Create criteria for evaluating the AI systems in relation to data privacy, biases, transparency, human oversight, employee influence, liability of vendors, and risk.
A high-risk AI recruiting tool would be one that influences the hiring process, deals with sensitive information, and is not transparent about its recommendations.
Step 3: Create an AI Intake and Assessment Process
Have the HR team submit their use case before using or acquiring an AI tool. The intake process should capture the purpose of the system and expected business outcomes.
Prior to implementation of an AI employee attrition tool, HR must submit documentation about which employee data will be used, who will have access to the predictions and if it acts on the recommendations.
Step 4: Develop an Escalation and Incident Response Process
Set out how any stakeholder can raise issues regarding any AI decision made. This could be due to privacy incidents, discriminatory impact, erroneous recommendation, or any unauthorized use.
For instance, if employees complain about any scheduling AI that discriminates against one particular group, then HR could escalate this to the committee, which may suspend the algorithm until the analysis is completed.
The Employee Experience Stake
- Recognizing Bias that Impacts Work Outcomes
HR knows what it means when biased outputs become part of hiring, promotions, pay, or development. Responsible AI in HR is enhanced through the relationship between model risks and employee impacts.
For instance, if an internal mobility tool is recommending employees for leadership positions, HR can investigate what data are driving those recommendations.
- Evaluating Whether the Use of AI is Compatible with the Organization’s Culture
The HR knows how employees would perceive the situation based on existing workplace culture and the level of trust. This helps determine whether an AI application is appropriate for the organization.
An organization with expectations around employee autonomy may need stricter controls before deploying AI tools that analyze employee activity.
- Measuring Experience Alongside Technical Performance
An AI system can meet accuracy while still damaging the candidate or employee experience. HR must incorporate complaints, adoption, and escalation data into the governance process.
In case an AI recruitment tool has improved screening effectiveness, but candidate complaints about their rejections have also increased, then a governance committee must look at their experience first.
What Organizations Do Right When HR Is at the AI Governance Table
This does not mean that HR has ownership of all decisions relating to AI. Rather, this means that both HR and IT need to be jointly responsible for the application of AI throughout the employee life cycle.
Paramita Patra is a content writer and strategist with over five years of experience in crafting articles, social media, and thought leadership content. Before content, she spent five years across BFSI and marketing agencies, giving her a blend of industry knowledge and audience-centric storytelling.
When she’s not researching market trends , you’ll find her travelling or reading a good book with strong coffee. She believes the best insights often come from stepping out, whether that’s 10,000 kilometers away or between the pages of a novel.
