Summary

This is a talk about so-called "artificial intelligence" (AI) in research, and pitfalls to avoid as it relates to labour and knowledge production. 

Key takeaways: 

1. Words matter - "AI" is intentionally ambiguous and having clear definitions for terms is necessary but insufficient; 
2. We should adopt an outcomes-based approach to thinking about AI issues; 
3. With the understanding that AI is (very often) not the problem. If we focus only on AI, then we risk making underlying problems worse! 

In addition to the editable slides here, the list of references, transcript, additional resources, and notes are published here: 

https://write.as/naclscrg/talk-ai-is-not-the-problem

and here: 

https://write.as/naclscrg/talk-ai-is-not-the-problem-follow-up

Files in this deposit

15 minute version

Hsing Open Science and Societal Impact talk 2024-04-25.pptx - A short, 15 minute, version of this talk given at the AESIS Open Science & Societal Impact conference 2024. The video recording is on the Internet Archive here: https://archive.org/details/AI-is-not-the-problem-2024-04-25

Hsing reproducibility symposium talk 2024-06-26.pptx - A tweaked version that relates to scientific reproducibility at the Reproducibility by Design symposium at the University of Bristol on 26 June 2024. The video recording is on YouTube here: https://www.youtube.com/watch?v=W_xYkQc_So8

30 minute version

Hsing TARG lab meeting talk 2024-11-22.pptx - A 30 minute version of this talk given at the University of Bristol TARG research group lab meeting on 22 November 2024 which expands on many of the themes explored in earlier versions. 

The video recording of this talk can be viewed on the Internet Archive: 

https://archive.org/details/AI-is-not-the-problem-2024-11-22