Automation without trust is expensive – why HR must co-lead AI governance

AI-driven automation is no longer a future concept. It is a present-day operational reality. Across industries, AI is reshaping how work is performed, how decisions are made, and how value is created. According to the World Economic Forum, 78 % of companies worldwide reported using AI in 2024, and adoption continues to accelerate. HR functions are no exception: the share of organisations deploying AI in HR processes grew from 26% in 2024 to 43% in 2025.

Yet employee confidence has not kept pace, with 61 % of employees reporting that they do not trust AI-driven decisions when those decisions lack clear human oversight. This trust gap was explicitly highlighted at the 2026 World Economic Forum meeting in Davos, where leaders emphasised that the primary constraint on AI’s impact is no longer technical capability, but responsible deployment and trust.

The conclusion is increasingly clear: automation without trust is not efficiency. It is hidden cost.

The hidden costs of trust-free AI
Organisations that deploy AI without strong governance expose themselves to operational, cultural, and legal risks. Initiatives built on weak foundations, poor data quality, unclear accountability, or opaque algorithms frequently underperform or fail. What begins as transformation can quickly become an expensive experiment.

When AI operates as a black box, errors are hard to detect and responsibility becomes diffuse. Employees confronted with unexplained decisions disengage or resist. Research cited by CIO.com indicates that nearly one third of employees admit to undermining AI initiatives, for example by bypassing official systems or using unapproved tools. These behaviours are rarely malicious; they are signals of broken trust.

The consequences ripple outward. Productivity declines, training investments fail, and ‘shadow AI’ proliferates, increasing cybersecurity, privacy, and compliance risks. At the same time, fear of job displacement erodes morale and accelerates resistance to change. Poorly governed AI can also create legal exposure, particularly when algorithms influence hiring, promotion or termination decisions. Talent attrition and reputational damage often follow.

These outcomes are not primarily technology failures. They are failures of leadership and change management.

Why HR must co-lead AI governance
Avoiding these pitfalls requires reframing AI implementation as a socio-technical transformation, not a purely technical project. This is why HR must play a central role in AI governance.

Leading organisations are moving away from IT-only ownership toward cross-functional governance models that include HR, IT, legal, compliance, and risk. While IT builds and deploys AI, HR understands its human impact: culture, ethics, workforce dynamics, and employee expectations.

HR increasingly helps define ethical standards, ensure algorithmic fairness, and assess whether AI tools align with organisational values. New roles such as Ethical AI Officers or HR AI Compliance Leads reflect this shift. AI governance is becoming a core leadership responsibility.

Most importantly, HR acts as the bridge between AI strategy and employee perception. Employees need to see AI as a tool that supports their work, not a hidden system that evaluates or replaces them without explanation. HR ensures organisations ask not only can we automate this? but Should we?

Building trust through transparency and capability
Trust in AI must be intentionally designed. Education is the starting point. OECD research shows that organisations investing in AI literacy and employee consultation achieve higher adoption and better outcomes. AI understanding must extend beyond technical teams to managers and frontline employees, who act as interpreters and change agents.

Transparent communication is equally critical. Employees should know where AI is used, why it is used, and where its limits lie. Initiatives explaining how algorithms work and when humans intervene have been shown to significantly increase adoption and reduce fear.

Nowhere is transparency more important than in HR-related AI decisions. Algorithms affecting recruitment or performance must be explainable and auditable. Responsible organisations define unacceptable uses of AI, implement human in the loop mechanisms, and create safe channels to challenge AI outcomes. These practices send a clear signal: AI is supervised, accountable, and designed to serve people.

Trust as the foundation of AI transformation
Globally, the push for trustworthy AI is accelerating. Frameworks such as the OECD’s AI Risk and Safety Framework underscore a key insight: responsibility is not a constraint on innovation, but a condition for scaling it. Regulatory scrutiny of workplace AI is increasing, particularly around transparency and bias.

For HR leaders, the mandate is clear. AI governance must be built on human oversight, accountability, and continuous learning. Employees must be engaged, not bypassed. Clear rules must define what AI can and cannot do.

In the age of intelligent automation, technology alone does not determine success. People do. When HR co-leads AI governance, automation becomes not a threat, but a sustainable strategic asset, one that enables organisations to innovate while remaining inclusive, resilient and profoundly human.

Milena Kowalik-Szeruga:
Milena Kowalik-Szeruga has been with Crowe Poland for several years, bringing experience in managing multidisciplinary projects as well as expertise in ESG and sustainable development. On a daily basis, she leads the ESG, Business Consulting, and Security Consulting teams and develops innovations for the business.

Milena is a graduate of Wrocław University of Science and Technology (Master’s degree in Environmental Engineering) and the University of Environmental and Life Sciences (Safety Engineering), as well as a long time auditor. She holds the Oxford AI Ethics, Regulation and Compliance certificate and the ESG and Sustainable Financial Strategy certificate from Saïd Business School, University of Oxford.