Is generative AI your new therapist? A cross-cultural model of generative AI affordances, trust, use intentions, and willingness to attend in-person therapy

Generative artificial intelligence (AI) tools are becoming an increasingly important source of mental health advice and support. This cross-sectional survey ( N = 1,011) conducted in Austria and Japan examined how perceived affordances of generative AI predict trust in generative AI and intentions to use it specifically in the context of mental health, and how these relate to stigma toward in-person therapy and willingness to seek professional help. The perceived affordances accessibility, personalization, and anthropomorphism consistently predicted both trust and use intentions across countries, whereas anonymity predicted trust only in Austria. Trust significantly predicted use intentions, and in Japan, this relationship was stronger among participants with higher perceived stigma, which also directly and positively predicted AI use intentions. Stigma negatively predicted willingness to attend in-person therapy in Austria but not in Japan. Finally, intentions to use generative AI for mental health were positively associated with in-person therapy willingness, suggesting a complementary rather than substitutive role of generative AI in mental health support. • Cross-cultural survey examining generative AI use for mental health support • Generative AI affordances predict trust and mental health AI use intentions • Anonymity predicts generative AI trust in Austria but not Japan • Stigma strengthens the link between trust and generative AI use intentions in Japan • Mental health AI use relates positively to in-person therapy willingness