received a notification that my Nature piece on Open Source AI is now up online!
ironically, however, it is not open access, and is closed behind a paywall. I've emailed to point out this irony, and am hoping they see it too :)
Remote
David Gray Widder
@davidthewid@hci.social
mastodon 4.7.3forever trying to help computer people think more critically about computer
Post Doc at Cornell Tech, PhD from CMU, exNASA/MSR/IntelLabs
art:IG@davidthewid
he/him
0 Followers
0 Following
29 Posts
Joined October 28, 2022
website:
🆕: Why AI ethics is someone else's job:
I'm studying how US undergrad CS majors choose careers.
If you teach or work with US based CS undergrads, I would be REALLY appreciative if you would encourage them to take this survey:
https://queensu.qualtrics.com/jfe/form/SV_czJcqzE8be8iRrE?d=sm
@AndrzejWasowski@social.itu.dk @clegoues@lebrady.net
Thanks for engaging with our work!! Glad you found ut useful, and hope you're doing well!
@wilbowma@types.pl this is what the youth call malicious compliance
Replying to
IT’S NOT ABOUT SPOTTING ISSUES, IT’S ABOUT HAVING *POWER* TO RESOLVE THEM
Even relatively powerful software engineers lack power to question the raison d’être of their company.
Replying to
I teach students to think critically about tech across large undergrad lectures, small PhD seminars, and reading groups and short courses for working adults.
I can teach (Critical) Data Science and HCI, and I revamp past courses to integrate critical concepts from STS & beyond.
Replying to
So far, my favorite part of my job has been mentoring 9 PhD students and undergrads!
Arriving from diverse backgrounds, I help them develop and achieve their own goals, including publishing in top venues and entering top CS and InfoSci PhD programs.
Replying to
I love interdisciplinary work, with collabs across Anthropology ( @dawnnafus, Lucy Suchman, Tamara Kneese), History (@histoftech), Law (@jtlg), Philosophy (Helen Nissenbaum), NLP, and Policy.
Replying to
Still, *Precarity* left many feeling unable to raise concerns, and unable to leave.
One said:
"Precarity will make your weigh your ethics less, right? … having a family, having dependents … medical conditions … I don’t really have a choice.”
Replying to
@WebCoder49 I don't know if I have the legal right to do that but that hasn't stopped me before lol so here:
https://davidwidder.me/open-nature.pdf
Replying to
Replying to
We also suggest that because using AI is a design choice, a focus on "*AI* ethics" elides wider scrutiny of incentives and business models, reifying what
Greene, Hoffman and Stark found in their "Better Nicer" paper.
Replying to
I'm also extremely proud to say that this is joint work with Derrick Zhen, one of my summer undergraduate research mentees in CMU's REUSE program.
Replying to
As #workers, they felt raising these kind of concerns about their employer's raison d’être was inherently futile:
As one said: "'Hey, either you shut down or I’m leaving’, and they’ll be like, ‘Alright, leave, I guess.’ It’s not really a concern you can raise."
Replying to
Raising concerns was often terrifying, as one said:
"it terrified me to confront an ‘authority’ figure, especially one who was the source of my financial well-being"
"I was so distraught over what I was being asked to do, I threw up in the parking lot before going into work."
Replying to
Some raised immigration precarity in particular, as I discussed with Chloe Xiang for
VICE News.
"H1Bs [often] need to find [a job] within a very short period of time or … leave the country … when that happens, you end up taking whatever is available.”
https://www.vice.com/en/article/n7z5px/twitter-employees-on-visas-cant-just-quit?utm_source=motherboard_twitter
Replying to
Some raise issues hoping to improve their product, for example making a VR system support multiple skin types.
But even then, they're dismissed as too costly, or something the client didn't ask for.
Replying to
But many respondents had larger concerns with their companies' business model or their industry's incentives.
One saw their own work as unproblematic, but pointed out: "The issue was more than the company as a whole, the business model [...] It was, you know, payday lending"
Replying to
And many did leave. Some refused, some slow walked, but many resigned and found different jobs.
As one said: "This work doesn’t get done without us".
But others worried they would be replaced, and the concern left unaddressed.
Replying to
We detail what happens when tech workers develop ethical concerns that challenge their employer's incentives or business models, where "Fair, Accountable, Transparent" interventions to help design more ethical products aren't helpful.
Replying to
My 16 peer-reviewed publications—appearing in venues from FAccT (x4), to Big Data & Society, to Nature—are on 11 syllabi from InfoSci to Gender Studies, cited in regulatory comments by Mozilla to the ACLU, and discussed in Wired and the New York Times:
https://www.nytimes.com/2024/05/17/business/what-is-openwashing-ai.html
Replying to
listen to the end to hear me talk about bombing data centers :)
Replying to
@travislloyd@hci.social I hate how people are moving from one centralized platform controlled by an egomaniac billionaire (Twitter/Musk) to another (Threads/Zuck).
Have we learned nothing?