00:19:23	Peter Mawhorter (he/him):	THANK YOU for doing a land acknowledgement!
00:27:34	Diana Tosca:	On a similar note, I’m wondering how we can improve communication between data scientist/those who work with the data and the public (those who the data actually, tangibly impacts). It seems like now there’s a big disconnect between how the data is used/processed and what the general public understands
00:33:29	Andrew Kun:	Questions - please type them here
00:38:21	Sadeghian, Shadan, Dr.-Ing.:	was there also a discussion of cooperation instead of replacement in this context? that “My part” would be done through a human-AI team collaboration?
00:39:00	Peter Mawhorter (he/him):	As a sometimes data scientist whose father is a Physicist (a “real” scientist) I’ve seen critiques of “data science” as a whole from very fundamental grounds: first, most data science is merely about correlations and never even attempts to look at causation, and second and deeper, most “real” scientists only develop hypotheses form theory before seeing the data instead of looking at the data to develop hypotheses, which is fertile ground for confirmation bias. To what degree do you think this kind of thing is really a deep flaw in data science vs. something that can be guarded against with the right strategies?
00:49:10	Vahid Yazdan-Panah:	Judea Pearl summarizes some of the fundamental limitations here. It was eye-opening for me (I suggest reading the last sentence in the paper) https://arxiv.org/pdf/1801.04016.pdf
00:50:25	Peter Mawhorter (he/him):	Thanks Vahid, that’s a great link!
00:53:24	Vahid Yazdan-Panah:	Welcome Peter, in Figure one (actually should be table 1), the claim is that with data, we can do association analysis but not interventions or studying counterfactuals.  
00:56:45	Diana Tosca:	https://www.ibm.com/blogs/research/2018/09/ai-fairness-360/
00:57:33	Nadia Fereydooni:	There is also a disconnect between data scientists and the those who generate the data; so the labor and social relations that created the data are removed and data is commoditized
00:57:41	Nadia Fereydooni:	https://dl.acm.org/doi/10.1145/1958824.1958906
00:57:44	Diana Tosca:	^yes!
00:58:17	Andrew Kun:	^^ just started reading Surveillance Capitalism...
01:01:47	Albrecht Schmidt:	is personalization (based on AI) the root cause of much of this bias issues?