AI Anxiety Index (AAI) Dataset: YouTube Consumer Comments on Artificial Intelligence — Emotional Response Corpus (n=944, April 2026)

This dataset contains 944 YouTube comments collected via the YouTube Data API v3 (April 2026) across five search domains targeting consumer emotional responses to AI systems:(1) AI anxiety/fear/privacy, (2) ChatGPT trust/distrust,(3) Algorithm unfairness/frustration, (4) AI personalization/surveillance, (5) AI identity threat/replacement. The corpus is the primary empirical validation dataset forthe AI Anxiety Index (AAI) — a novel computational metricmeasuring the proportion of anxiety markers relative tototal emotional markers (anxiety + trust + curiosity +irritation) in consumer-generated text. KEY FINDINGS:- Mean AAI: 0.5384 (anxiety dominant over trust/curiosity)- Anxiety marker density: 0.8356 per 100 tokens- Curiosity marker density: 0.5734 per 100 tokens- Trust marker density: 0.1313 per 100 tokens- IAEM Preservation cluster dominant: 148/248 (59.7%)- Most anxious domain: AI privacy/fear (AAI=0.5923)- Most balanced domain: Algorithm unfair (AAI=0.3333) The AAI operationalises the Identity-Aligned EmotionalModel (IAEM) computationally, extending Lazarus (1991)cognitive appraisal theory and Oyserman (2015)identity-based motivation theory to AI consumer contexts. This work extends prior DBA research on AI and Greekconsumer behaviour (citation omitted blind review) witha new empirical instrument grounded in real consumer data. Files:- ai_anxiety_videos.csv: 50 unique videos metadata- ai_anxiety_comments.csv: 944 raw comments- ai_anxiety_aai_results.csv: AAI scores + IAEM clusters Model: YouTube Data API v3. Analysis: Python 3.12,Google Colaboratory, April 2026.