A safe place to ask? Identity, privacy, and checking when LGBTQIA+ adults use AI for climate information in Nigeria’s South-South

Abstract

In Nigeria's flood-prone South-South, lesbian, gay, bisexual, transgender, queer and intersex and asexual (LGBTQIA+) adults are more frequently relying on general-purpose artificial intelligence (AI) tools to decipher climate information. This study on 502 LGBTQIA+ AI users in Calabar, Uyo, Warri, and Yenagoa used the typologies of minority stress theory and Communication Privacy Management theory to explore the sources of their climate information, the motives for using AI, whether they verified information from AI, and who they would rely on for information. More than half of the information AI received was via WhatsApp (72.3%), and 42.8% cited no official agency. Almost all respondents in local language used English, Pidgin or a combination of both. Weaker reporting of increased safety reasons for using AI by transgender/intersex, disabled and displaced participants. There was no significant difference in OR for the relationship between privacy worry and checking while there was a significant OR for the relationship between digital skills and checking. Interestingly, without official sources, participants did not appear to be less likely to accept official messages; trust in AI was the strongest predictor of adherence to AI. Warning agencies should post verified multilingual messages via WhatsApp and community organisations should develop checking skills and not rely on privacy warnings.