- When: October 27th, November 10th, 23rd, December 1st (4 days, Each week)
- Where: Online https://bit.ly/Fall-in-HCI-at-KAIST
This year’s HCI@KAIST fall colloquium invited four speakers from diverse HCI domains.
Sherry Tongshuang Wu from Carnegie Mellon University
Practival AI Systems and Effective Human-AI Collaboration
As AI systems (such as LLMs) rapidly advance, they can now perform tasks that were once exclusive to humans. This trend indicates a shift towards extensive collaboration with LLMs, where humans delegate tasks to them while focusing on higher-level skills unique to their capabilities. However, haphazard pairing of humans and AIs can lead to negative consequences, such as blind trust in incorrect AI outputs and decreased human productivity. In this talk, I will discuss our effort in promoting effective human-AI collaboration, by ensuring competence in both humans and AIs for their respective responsibilities and enhancing their collaboration. I will cover three themes: (1) Evaluating LLMs on specific usage scenarios; (2) Building task-specific interactions that maximize LLM usabilities; and (3) Training and guiding humans to optimize their collaboration skills with AI systems. In my final remarks, I will reflect on how AI advances can be viewed through the lens of their usefulness to actual human users.
Sang Won Lee from Virginia Tech
Toward Computuer-mediated Empathy
This talk discusses ways to design computational systems that facilitate empathic communication and collaboration in various domains. In contrast to using technologies to develop users’ empathy for targets, I emphasize the duality of empathy and highlight empowering targets to express, reveal, and reflect on themselves. An ongoing framework will be introduced, and I will focus on recent projects that explore sharing perspectives, self-expression, and self-reflection as a means to mediate empathy in interactive systems from target perspectives.
Janghee Cho from National University of Singapore
Design for Sustainable Life in the Work-From-Home Era
Navigating the complexities of the contemporary human experience is precarious, marked by latent but pervasive anxiety and uncertainty. In this talk, I draw on a reflective design approach that emphasizes the value of human agency and meaning-making processes to discuss design implications for technologies that could help people (re)establish a sense of normalcy in their everyday lives. Specifically, the focus centers on recent projects that investigate the role of data-driven technology in addressing well-being issues within remote and hybrid work settings, where individuals grapple with blurred boundaries between home and work.
Audrey Desjardins from University of Washington
Data Imaginaries
0s and 1s on a screen. The Cloud. Fast moving. Clean. Efficient. Exponentially growing. Data Centers. Code on the black screen of a terminal window. Buzzing. Such images construct part of commonly shared imaginaries around data. As data increasingly become part of the most intimate parts of people’s lives (often at home), it remains a largely invisible phenomenon. In particular, one of the leading challenges currently facing Internet of Things (IoT) is algorithmic transparency and accountability with regards to how IoT data are collected, what is inferred and who they are shared with. From a home dweller’s perspective, data may be available for review and reflection via graphs, spreadsheets, and dashboards (if at all available!).In this talk, I instead argue for other modes of encountering IoT data: ways that are creative, critical, subtle, performative, and at times analog or fictional. By translating data into ceramic artifacts, performance and interactive installation experiments, fiction stories, imagined sounds, faded fabric, and even data cookies, I show a diversity of approaches for engaging data that might capture people’s attention and imagination. As a result, this work uncovers ways to make data more real, showing its messiness and complexities, and opens questions about how data might be interpreted, and by whom.