I've been wrestling with the gap between HCI curricula, which tend to focus on the design process, and the big theories and ideas that animate us in HCI. Here's a model that has starting working well at @StanfordHCI:
Despite the promise of AIs improving human decision making, a frustratingly resilient research result has found that people with AI decision-making aids are no better than people alone or AIs alone. Why? Overreliance: we take the AI's recommendations even when we shouldn't.
Congratulations to @joon_s_pk and team on the #UIST2023 Best Paper Award for Generative Agents!
Happily, we already know that the agents can throw a party in celebration.
The #UIST2023 best paper awards and honorable mentions have been announced: programs.sigchi.org/uist/2023/awar…
Congrats to all the authors and thanks to Andy Wilson, who served as Awards Chair. Excited to see these papers and more next week - last chance to register at Standard Rate!
What unifies, and what distinguishes, social media designs? Are all the Twitter spinoffs actually meaningfully different designs from each other? Form-From is a design space from @amyxzh, myself, @karger, and @answergarden that will appear at #cscw2024arxiv.org/abs/2402.05388
They kept saying that I looked like an “older, retired Tom Holland”. And now my PhD students have celebrated my birthday with the most on-brand custom bobblehead possible.
Fully 80% of papers at UIST from 1990-2010 have been cited by US patents. 20% overall at CHI+CSCW+ Ubicomp+UIST; 13% of papers at all @sigchi sponsored venues.
AAAI/IJCAI are 5%, ACL/EMNLP/NAACL are 11%, CVPR/ICCV/ECCV are 25%.
What’s the industrial impact of human-computer interaction (#HCI) research? Does HCI research contribute to technological inventions and products? Or are most of its insights ignored by the industry? Our #CHI2023 paper provides new evidence for these long-standing questions.
People using explainable AIs are empirically no better at decision making than people working alone. @HelenasResearch, an Honorable Mention at #CSCW2023, demonstrates that this failure is because AI explanations typically require lots of cognitive effort to verify.
Today's social media AIs encode values—can we mitigate societal harms by making these values explicit and tuneable? Excited to share our #CSCW24 paper introducing societal objective functions, which translate social science constructs into algorithmic objectives for social media!