Recent data from Deloitte’s TrustID Index shows a sharp drop in employee trust toward employer-provided AI tools. Between May and July 2025, trust in generative AI declined 31%, while trust in agentic AI, systems that act independently, fell 89%. Despite AI’s potential to ease workloads, frontline workers increasingly turn to unapproved tools, signaling not a rejection of AI itself but a lack f confidence in company-sanctioned systems. 

According to Harvard Business Review, the reasons are both practical and psychological. Employees report feeling that AI tools are imposed without input, and worry that adoption may eventually displace them. To address this, organizations are adopting several interventions. Deloitte recommends measuring trust through behavioral metrics that assess transparency, reliability, capability, and humanity. When these metrics are included in broader performance evaluations, they offer actionable insight into where AI adoption may be faltering. 

Investment in skills is equally important. Deloitte found that employees who received hands-on AI training reported 144% higher trust in those systems than their untrained peers. This trust correlates with greater motivation to upskill and higher rates of engagement. Co-creating tools with employees, rather than designing them in isolation, has also improved uptake. 

Walmart and Colgate-Palmolive have introduced programs that integrate employee feedback directly into AI development and deployment. These models highlight a growing understanding: building trust in AI requires consistent investment in human-centered design and clear communication.