AEDI v3 Dataset — Cross-Cultural AI Emotional Discourse Corpus: Native-Language Google Trends Across European Anglophone, Romance, and Germanic Linguistic Cultures (2022-2026)

The AEDI v3 (AI Emotional Duality Index, version 3) corpus measures the cross-cultural balance between AI Fear and AI Curiosity in naturalistic European search behaviour over 220 weeks (January 2022 to April 2026), using native-language Google Trends queries matched to Hofstede's (2001) Uncertainty Avoidance Index (UAI) values for each market. The corpus addresses a methodological limitation of prior English-only computational discourse research on AI: the exclusion of substantial non-anglophone European AI publics whose fear-curiosity discourse occurs in their native linguistic registers rather than in English-language search behaviour. Five fear/curiosity dimensions are operationalised through paired queries in each of ten European languages: AI Job Threat, AI Surveillance, AI Capability, AI Future, and AI Trust. Each dimension captures a distinct DEET (Dual Emotional Engagement Theory) phase grounded in established theoretical foundations: Loewenstein's (1994) Information Gap Theory of curiosity; Rogers's (1975) Protection Motivation Theory; Cave, Coughlan, and Dihal's (2019) analysis of AI fear narratives; Litman's (2005) epistemic curiosity; and Zlotowski, Yogeeswaran, and Bartneck's (2017) autonomy threat framework. Native-language query construction follows standard cross-cultural research conventions for Google Trends measurement across non-anglophone markets. The dataset comprises: - aedi_v3_summary.csv: country-dimension aggregate scores with UAI values, regional classification (Southern/Northern Europe), and validity counts - aedi_v3_weekly.csv: 4,014 weekly observations in long format with fear_search, curiosity_search, AEDI ratio, dimension, country, language, region The principal empirical finding is a Pearson correlation r = 0.83 between Hofstede UAI and country-level mean AEDI across the six markets with sufficient native-language search volume: Italy (UAI=75, AEDI=0.72), Netherlands (UAI=53, AEDI=0.67), Finland (UAI=59, AEDI=0.67), Germany (UAI=65, AEDI=0.40), Sweden (UAI=29, AEDI=0.25), and Denmark (UAI=23, AEDI=0.17). This correlation provides cross-cultural empirical support for the prediction that high-UAI cultures generate fear-dominant AI discourse in naturalistic search behaviour, extending earlier qualitative work on Greek consumer AI fear (citation omitted blind review) to a multi-country comparative framework. Additional dimension-level findings document that AI Capability discourse is most fear-dominant (mean AEDI = 0.92) while AI Trust discourse is most curiosity-dominant (mean AEDI = 0.21), indicating that frame-dependence operates at the discourse level within national markets in addition to the cross-cultural level between them.