Kars Alfrink 🟥
@kars.leapfrog.nl
290 followers 200 following 370 posts
Designer turned design researcher. Postdoc at TU Delft. Exploring contestable AI. www.contestable.ai
Posts Media Videos Starter Packs
kars.leapfrog.nl
It is a teaser of sorts for a position paper I have been writing since earlier this year in the semi-open. I hope to do more work on this topic in the years ahead. So it was very productive to discuss this at the symposium.
kars.leapfrog.nl
My contribution explored how we can use design to enable community control over AI systems development towards prosocial ends.
kars.leapfrog.nl
Talks ranged from arguments for reform to abolition of "AI" (scare-quotes intentional) and a variety of positions in between. The ones I enjoyed the most combined a solid grounding in technology with sophisticated political thinking. I learned a lot!
kars.leapfrog.nl
I had a great time presenting at the Politics of AI symposium at Goldsmiths. Many thanks to Dan McQuillan, Fieke Jansen, Jo Lindsay Walton, and Pat Brody for inviting me and putting together such a thought-provoking program.
kars.leapfrog.nl
Excited to share my People's Compute research project ideas at the Politics of AI symposium at Goldsmiths, organized by @danmcquillan.bsky.social and others, this Thursday. www.criticalinfralab.net/preview/?id=...
critical infrastructure lab » preview
www.criticalinfralab.net
kars.leapfrog.nl
Full update with all the details, talks, papers, and photos: leapfrog.nl/blog/archive...
kars.leapfrog.nl
The din of AI hype and criti-hype continues online. I keep trying to stay grounded and navigate between naive embrace and cynical rejection. Not easy, but it is the work.
kars.leapfrog.nl
Also spent some time in the streets protesting €2 billion in education cuts. Because apparently that's what it takes to defend the future of research and education in the Netherlands these days.
kars.leapfrog.nl
Published research on contestability in AI systems, developed a new master's course on responsible AI design, and coached four thesis students through graduation.
kars.leapfrog.nl
One year into my postdoc at Delft University of Technology, not sure where the time went. Quite the year. 🧵
kars.leapfrog.nl
Same here. I’ve been enjoying NTS.
kars.leapfrog.nl
Some more details on Ameya's thesis and the other finalists here: www.delftdesignforvalues.nl/what-we-do/a...
kars.leapfrog.nl
Also, a shoutout to master's student Ameya Sawant, whose thesis was one of three works selected for the thesis award. It continues to be a pleasure to mentor this very interesting project on designer autonomy in times of AI.
kars.leapfrog.nl
I am also very pleased to have the white paper I contributed to out in the world. I wrote a short piece on designing AI for autonomy. Udo Pesch did a great job editing it all together.
kars.leapfrog.nl
Human autonomy was the driving value behind my PhD research on contestable AI, and continues to be of great interest to me. So I thoroughly enjoyed attending the Design for Human Autonomy conference.
kars.leapfrog.nl
About AI ethics, not design ethics, but still related:
> van Maanen, G. (2022). AI Ethics, Ethics Washing, and the Need to Politicize Data Ethics. Digital Society, 1(2), 9. doi.org/10/g833qb
kars.leapfrog.nl
Drake hotline bling meme "design ethics" / "design politics".
kars.leapfrog.nl
Many systems labeled as "AI" simply don't work. We should demand evidence of the functionality of specific systems before debating their ethics (cf. Raji et al., 2022).
kars.leapfrog.nl
The challenge, as I see it, is to do criticism of "AI" while not falling into the trap of reinforcing its "uncontroversial thingness" (Suchman, 2023).
kars.leapfrog.nl
Understood as a floating signifier, though, saying "AI" is or is not useful makes no sense.