AI and Workplace Wellbeing | Micah Kaats and George Ward | University of Oxford

Value of the Information & Strength of the Argument

The presentation provides valuable insights into the underexplored area of AI’s impact on job quality and wellbeing. The argumentation is solid, building on a comprehensive review of existing literature and using multiple large-scale datasets. The authors carefully validate their measures and control for potential confounders like wages and education. They also acknowledge limitations, such as the correlational nature of the findings and the gap between exposure and actual use. The discussion of different AI usage patterns adds nuance, though the mixed results for usage categories suggest the need for further investigation.

Scientific Rigor, Source Quality, Title Accuracy

The research demonstrates high scientific rigor, with a clear methodology and reliance on reputable data sources (Indeed, ONET, Anthropic). The authors validate their AI exposure measure against existing ones and discuss potential biases. The title accurately reflects the content, and the presentation is well-structured. The use of multiple studies strengthens the overall argument. However, the lack of peer review and the preliminary nature of the results temper the overall assessment.

The title accurately reflects the content, which focuses on the relationship between AI and workplace wellbeing.

8/10

The presentation is based on rigorous research using large-scale datasets (Indeed, ONET, Anthropic) and validated measures. The authors are affiliated with prestigious institutions (Harvard, INSEAD, Oxford). The methodology is transparent, and limitations are acknowledged. However, the findings are preliminary and not yet peer-reviewed.

• 00:00Introduction
• 03:56Motivation
• 06:08AI exposure effects
• 16:40AI use effects
• 22:59Individual effects

• Indeed — Crowdsourced workplace wellbeing data
• ONET — Occupational data from the Bureau of Labor Statistics
• Anthropic — Data on Claude usage
• Frey & Osborne (2017) — Susceptibility to computerization measure
• Webb (2019) — Exposure to robotics, software, and AI
• Eloundou et al. (2024) — GPT-4 exposure measure

• Eloundou et al. (2024) — Similar positive association between AI exposure and wellbeing
• Webb (2019) — Modern AI exposure positively associated with wellbeing

• Frey & Osborne (2017) — Older automation exposure measures show negative association with wellbeing