Ph.D Student at Gatech @dmgatech | Prev. @upstageai @hcikaist @snuitct | She/Her
Atlanta, USA
Joined September 2020
- Tweets90
- Following167
- Followers113
- Likes274
Pinned Tweet
I and @richmondywong are excited to present our CHI 2025 paper next Tuesday in Yokohama! While AI adoption is often framed as a linear and inevitable process, our work critically examines the moments when AI was not adopted or was discontinued through AI "non-use."
It was great to work with NLP researchers! If you are attending EMNLP 2025 this year, please stop by our poster :)
Thank you for highlighting our work! I had such a fun conversation with @techsquareatl :)
Why are some UX designers saying no to AI? 🤔
Tech Square's Inha Cha (@inha423) and Richmond Wong delve into the reasons behind the non-use of AI in design, emphasizing the importance of human judgment, ethics, and context. Her research challenges assumptions around AI adoption.
Check it out! 🔗 buff.ly/JKNqrNQ
@gt_lmc @GeorgiaTech
We identify how factors such as organizational dynamics, tool constraints, ethical considerations, and team practices come together to shape AI use/non-use in practice.
📍 April 29th, 11:10–11:22am @ Room G304
programs.sigchi.org/chi/2025…
🔥 Paper Link (available by May 25th): doi.org/10.1145/3706598.3713…
If you’ll be at CHI, I’d love for you to join us in rethinking what AI usage entails. Rather than assuming adoption as the default or ideal outcome, we consider how non-use can be a thoughtful, even necessary response—especially when AI may be unnecessary or harmful.
Inha Cha retweeted
[NEW PAPER ALERT!] In this work, we present PROFILE, a framework designed to discern the alignment of LLM-generated responses with human preferences at a fine-grained level (length, formality and intent etc). Our key finding is a significant misalignment between LLM's output and human preferences. LLMs tend to over-prioritize response length across tasks, EVEN in summarization. On the contrary, LLMs's evaluation aligns better with human compared to generation. This introduces a huge improvement space under LLM's self-correction framework. Amazing experience working with KAIST students 👩👩👧👧 @juhyunohh @euns0o_kim @jiseon_kim1 and @inha423. Thanks to Prof. @aliceoh and @WilliamWangNLP.
Inha Cha retweeted
Find out more in our paper! This work is a collaborative effort with my amazing co-authors @euns0o_kim @jiseon_kim1 @WendaXu2 @inha423 @WilliamWangNLP @aliceoh !
📜: arxiv.org/pdf/2410.06965
A Telegram channel with 220,000 members created and shared deepfaked pornography images and videos—many victims and perpetrators were minors. What happened in the place I had been born, raised, and lived left me miserable. m.koreaherald.com/amp/view.p…
Deepfake sex crimes surge among teenagers in S. Korea
: Telegram chats distributing sexually exploitative deepfakes expand from universities to schools
chosun.com/english/national-…
How can I coexist with those who find joy in exploiting the vulnerable? I question whether the tech ethics and justice I advocate for truly can resonate in this world…
Inha Cha retweeted
I will present our #CVPR 2024 paper “LTM: Lightweight Textured Mesh Extraction and Refinement of Large Unbounded Scenes for Efficient Storage and Real-time Rendering" on Wed (June 19th) 5:00 to 6:30 PDT!
openaccess.thecvf.com/conten…
Inha Cha retweeted
📢 Recruiting participants for a paid study about LLMs
I'm looking for US-based adults for a 1 hour study on Zoom (payment: $15 Amazon gift card), especially those with technical knowledge of LLMs
Sign up at forms.gle/6aGfJNrfhpuQzmz59
Please RT and help out a grad student 😁
Excited to share our DIS'24 preprint!🎉 Together with @ajplai and @richmondywong, we developed "Ethics Pathways," a design activity that helps reflect on the nuanced engagement with infrastructure and processes to action ethics within institutions. arxiv.org/abs/2405.16654v1
This opens up new avenues for HCI to address ethics beyond individual decisions, considering social and institutional power structures. We look forward to present this work at DIS 2024!