A Trust Crisis in AI Implementation Exists as Organizations Rush to Deploy GenAI

Titled State of AI 2025: Mid-Year Report: Lack of Trusted Content Emerges as Achilles Heel, the study is based on a survey of 316 KM professionals, as announced on June 3, 2025 by eGain. An important finding is that executives expressed more confidence in AI than did frontline staff. Knowledge silos emerged as a key barrier to reliable AI output.

Joseph McKendrick, the report's author, emphasizes the need for integrated solutions: "Simply slapping GenAI over existing knowledge silos only takes chaos to the next level. Organizations need to unify their knowledge foundations with trusted hubs that keep human experts in the loop."

The study identifies fragmented knowledge systems as the primary culprit behind unreliable AI responses. A majority of organizations (55%) maintain three or more separate knowledge repositories, leading to the top barriers cited by respondents: erroneous answers (61%), inconsistent responses (61%), and non-compliant information (54%).

Despite trust concerns, customer service applications are proving most successful, with customer self-service leading adoption at 45%, followed by human-assisted contact center support at 38%.

Investment momentum remains strong, with 75% of organizations planning to increase generative AI budgets. However, 60% maintain separate AI and knowledge management budgets, potentially hampering integrated approaches. Additionally, 62% report being unsatisfied with their current knowledge management platforms.

The study found overwhelming agreement (94%) on the importance of human expert involvement in ensuring AI quality, suggesting successful AI implementation requires careful orchestration of human and machine intelligence.

About the Study: State of AI 2025: Mid-Year Report: Lack of Trusted Content Emerges as Achilles Heel was conducted in spring 2025 among KMWorld magazine readers, representing KM professionals across multiple industries. To read the full report, go here