Ethics in engineering
Responsible research
Plagiarism
Harassment
Privacy
Dual-use research
Obligation to the public interest
Obligation to explain (in math, who decides what constitutes a proof)
Inclusion, the Framingham studies
Exploitation, war atrocities, Tuskegee syphilis study, protection of vulnerable populations
Mismeasure of Man (Stephen Jay Gould)
Colonialism, Guns, Germs and Steel (Jared Diamond)
Commercialism
Biases in rewards and recognition
Who decides which research to fund
Mythologies in research (e.g., “alpha male” wolves)
Gender biases in research (e.g., gender traditions in fields, feminist glaciology, predation/competition versus mutualism/symbiosis)
See work by Michael Barany (Edinburgh)
“What does it mean for a theorem to be biased? is an unintuitive question that becomes a lot more understandable when you think about some of the ways it can be biased.”
Thanks for this, Scott!
I think the key to the theorem-bias question is to start mapping out what goes into a theorem and what comes out of it in human terms, i.e. how is a theorem a part of a community of provers/knowers/learners/evaluators/etc? So bias in the results of pure mathematics is often about what people look for in a demonstration, what the theorem takes for granted, how it emphasizes some ways of knowing at the expense of others. There’s a key shift in perspective that I really work on with postgrad students to move away from seeing theorems and theories as isolated statements in papers, and rather start to see them as nodes in human communities with identifiable effects on those communities --- uneven effects, hence the possibility of identifying bias.
I wrote about this in the context of the Intermediate Value Theorem a while ago https://www.ams.org/notices/201310/rnoti-p1334.pdf , for example. With the “pure maths” postgrad students I tend to talk about this after they’ve read chapter 3 of Harris’s Mathematics Without Apologies, on the different justifications mathematicians and others have for why careers in pure mathematics should be possible. With my undergraduate history of science students, I like to use Mary Terrall’s chapter about gender and the French Académie des Sciences here https://www.google.co.uk/books/edition/The_Sciences_in_Enlightened_Europe/ttGgd6mec1MC?hl=en&gbpv=1&dq=terrall%20gender%20mathematics&pg=PA246&printsec=frontcover . It can also be helpful to contrast the outcomes of different historical claims to major theorems: Fermat’s Last Theorem, the Poincaré Conjecture, the unresolved ABC controversy, and some really interesting historical claims to the Riemann Hypothesis.
I’d read those Cambridge resources and Wikipedia site with a grain of salt, can talk about why.
I’ve been on and off in conversations with colleagues in Engineering at Edinburgh about, e.g., the aim to decolonise engineering education. Some interesting models of ethics in engineering degrees in the US that friends of mine have been involved with. The history of engineering education since 1800 at least helps explain why it’s so hard to change, but also what a big effect changes can produce.
I’d be glad to talk about these things more with you. Later next week or next month, perhaps, depending on your sense of urgency?
Best,
Michael
From: Ferson, Scott <Scott.Ferson@liverpool.ac.uk>
Sent: 17 February 2022 11:49
To: BARANY Michael <m.barany@ed.ac.uk>
Cc: Gray, Nicholas [nickgray] <Nicholas.Gray@liverpool.ac.uk>
Subject: ethics in engineering
Dear Michael:
Your presentation at the CDT directors’ meeting today was intriguing.
Most of what you mentioned seemed clear but I was interested when you said something like “‘What does it mean for a theorem to be biased?’ is an unintuitive question that becomes a lot more understandable when you think about some of the ways it can be biased.” I wonder whether you can give me a hint about what it means for a theorem to be biased. Can you suggest a reference?
I’ve just found https://ethics.maths.cam.ac.uk/assets/dp/18_1.pdf and even https://en.wikipedia.org/wiki/Ethics_in_mathematics which I will need to absorb.
I might mention that at our Risk CDT in Liverpool, we have a working group on what I think is a closely related subject we call “humane algorithms” (https://sites.google.com/site/humanealgorithms/, click on the three horizontal bars in the top, left corner to see the topics, many of which are under the Say tab). I’d be very interested in any thoughts you might have about it.
Our responsibilities to the undergraduates are perhaps even more weakly realised. In the seventh week of a 7.5-credit module that all engineers take in their first semester at uni, we squeeze in the animated slides in the attached file. Of course this is literally a bullet-point approach rather than something more engaging such as student-led discussions. Engineers seem to have come to the realisation of the centrality of ethics remarkably late considering how direct and consequential their work can be. The inertia that stands in the way of changing their curriculum is rather large. Do you have suggestions for addressing the important issues for this younger group?
Best regards,
Scott