AI in multilingual organizational communication: the impact of culturally adapting machine translation on global leadership self-efficacy

Purpose – This paper examines and explains the impact of machine translation (MT) tools with culturally adapting capabilities and without such capabilities on global leadership self-efficacy, motivation to learn and willingness to join the multinational organization (MNO). The perceptions of tools' agency and effectiveness are suggested and tested as the underlying cognitive mechanisms that explain the impact of MT tools. Design/methodology/approach – The research is based on three scenario-based studies conducted across different populations and languages. Participants were presented with a job description that involved the use of either culturally adapting or non-adapting MT tools and their levels of global leadership self-efficacy and behavioral intentions were measured. In addition, the perceptions of tools' agency and effectiveness were assessed. Quantitative analysis was employed to test the direct and indirect effects of MT’s capabilities on self-efficacy, learning motivation and behavioral intentions. In addition, the mediating role of MT tools' perceived agency and effectiveness was assessed. Findings – The findings demonstrate that access to culturally adapting MT significantly enhances global leadership self-efficacy compared to non-adapting MT tools. Higher self-efficacy positively correlates with the intention to join the global company. Contrary to our hypothesis, access to culturally adapting MT tools did not hinder participants’ motivation to learn about the foreign culture or language. Although culturally adapting MT is perceived as more effective and more agentic and autonomous, only the perceived effectiveness significantly mediates the impact of the tool on self-efficacy. Motivation to learn is affected by both perceptions in a competing manner: perceived tools' effectiveness has a positive impact on the motivation to learn, while perceived agency has a negative impact. Research limitations/implications – While this study highlights the impact of culturally adapting MT on global leadership self-efficacy, several limitations remain. First, it assumes high translation quality, though MT’s accuracy and ability to convey nonverbal cues remain uncertain. Second, it examines MT access rather than its long-term use and effects on learning motivation. Finally, the scenario-based experimental design limits real-world applicability. Future research should investigate MT performance, long-term implications and organizational consequences using field studies and qualitative methods. Practical implications – Cutting-edge MT can expand the talent pool by enabling individuals with lower language proficiency to participate in multicultural environments. Additionally, effective MT use can reshape language-related power dynamics and promote multilingual cooperation. However, managers must also offer training to prevent over-reliance on AI tools, ensuring continuous learning and communication skills development. Implementing AI-MC literacy programs will help employees use these tools effectively while mitigating risks of over-trust and miscommunication. Social implications – The findings suggest that AI-mediated communication tools, particularly culturally adapting MT, can enhance global self-efficacy and encourage participation in international work environments. However, the perception of AI as highly agentic may reduce motivation to learn new languages and cultures, fostering dependence on technology rather than personal skill development. This could widen knowledge gaps and create over-reliance on AI in cross-cultural communication. On the other hand, when AI is perceived as effective but not overly autonomous, it may support learning motivation. Policymakers and educators should balance AI integration with strategies that promote continuous learning and critical assessment of AI tools. Originality/value – This study uniquely integrates AI-mediated communication (AI-MC) into cross-cultural organizational contexts, combining the Theory of Planned Behavior and Agency and Social Cognition theory. It empirically demonstrates that culturally adapting MT enhances global self-efficacy while revealing the paradoxical effect of AI agency on learning motivation. Unlike previous research, it highlights the psychological trade-offs between AI effectiveness and agency, showing that while AI can empower users, excessive reliance may hinder personal development. These insights extend discussions on AI’s role in international business and education, emphasizing the need for careful AI implementation to balance efficiency with human skill cultivation.