YouTube SML Dataset — Public Sentiment Toward AI Tools in Education (VADER Analysis, March 2020 – March 2026)

This dataset contains 758 YouTube comments collected via the YouTube Data API v3 across five thematic categories related to artificial intelligence in education. Data collection covers the period March 2020 to March 2026 — a six-year window spanning the COVID-19 forced digitalisation of education (March 2020), the mainstream emergence of generative AI tools including ChatGPT (2022–2023), and the current phase of institutional response and ethical debate (2024–2026). Each comment was analysed using VADER sentiment analysis (Hutto & Gilbert, 2014), producing positive, negative, neutral, and compound scores. The five thematic categories are: (1) AI as Teaching Tool, (2) Educator Anxiety and Resistance, (3) Student Adoption and Academic Integrity, (4) Ethics, Surveillance, and Data Privacy, and (5) Sustainability and the Future of AI-Driven Education. VADER Results Summary: T1 AI as Teaching Tool: 46.7% Positive / 23.9% Negative / 29.4% Neutral — Compound: 0.14 (N=180) T2 Educator Anxiety: 60.8% Positive / 20.0% Negative / 19.2% Neutral — Compound: 0.24 (N=120) T3 Student Adoption & Integrity: 59.7% Positive / 17.8% Negative / 22.5% Neutral — Compound: 0.31 (N=129) T4 Ethics & Surveillance: 56.0% Positive / 29.8% Negative / 14.3% Neutral — Compound: 0.21 (N=84) T5 Sustainability & Future: 66.9% Positive / 6.1% Negative / 26.9% Neutral — Compound: 0.42 (N=245) Data collected and analysed as part of Master's thesis research, LEADS MSc programme in Learning, Digitalization, and Sustainability, Jönköping University, Sweden (Spring 2026). Python 3.13.3, vaderSentiment 3.3.2, Google API Python Client 2.193.0.