In AI We Trust: Opportunities and Threats of Using Generative Artificial Intelligence in Public Relations

Introduction
Despite the overwhelming adoption of generative artificial intelligence (AI) in the daily tasks of public relations (PR) practitioners, the consequences of relying on AI-generated content for publics’ trust towards organisations remain insufficiently explored (Yue et al., 2024). The present study addresses this research gap by investigating the components of trust toward organisations, with a specific focus on information-seeking behaviours and their epistemic consequences.
Literature Review
Galloway and Swiatek (2018) argue that public relations (PR) focuses primarily on human-to-human relationships. However, they also suggest that individuals can satisfy their need for meaningful connection through non-human entities, such as ideological movements, organisations, and brands. Consequently, members of the public exhibit information-seeking behaviours that require content to be credible, reliable, and transparent. While artificial intelligence (AI) can advance the PR industry by accelerating the analysis of large datasets, significant concerns remain regarding content quality (e.g., AI hallucinations), data management, and the evolving relationships between organisations and their publics (Cheng & Verčič, 2025). Recent industry challenges illustrate these risks, such as Amazon’s historical gender bias in its recruitment algorithms and the January 2026 deepfake controversy involving Grok, where the AI was used to generate explicit imagery via nudity prompts.
To address these challenges, Younger (2024) advocates for a conceptual shift, viewing generative AI (GenAI) as a stakeholder rather than merely a technology, and recommends that organisations develop specific AI engagement strategies. Beyond managing the technical limitations of GenAI, PR practitioners must understand the public's expectations of the technology and scrutinize the components of trust formed during the information-seeking process.
Method
The present study addressed the need to understand the antecedents of epistemic trust toward organisations that communicate using GenAI versus human-created content. Through an online experimental design, content authorship and organisational familiarity were manipulated. A press release from an international consumer goods organisation and an AI-generated version of the same release were used to examine how members of the public evaluated dimensions related to epistemic trust.
Results and Discussion
Results revealed significantly higher trust toward organisations that communicate using GenAI content, p = .041. A five-component model of trust for Gen Z consumer stakeholders (R²adj = .74) was proposed, consisting of perceptions of authenticity, transparency, predictability, reliability, and knowledgeability. Moreover, there was a significant difference in how the perception of authorship influenced the expected components of trust. While human-generated trust was influenced by the reliability and transparency of the organisation, GenAI trust was explained by authenticity and reliability. Although publics expect reliability from both human and AI-generated content, GenAI is paradoxically perceived as more authentic than human-created content. This study is limited to exploring the cognitive components of trust; its affective component requires additional investigation.
Practical Implications
The risks and benefits for organisations building trust with their publics through AI versus human-created communication are discussed. The results suggest that GenAI can support practitioners when communicating novel information, such as major innovations or the introduction of new products. However, AI scepticism may arise when GenAI is used to manage relationships between publics and familiar organisations with established reputations. To mitigate this negative effect, practitioners should produce and communicate human-centric content in a transparent manner while explicitly emphasising human authorship.