📢
#CHI2024
Paper: Using tools like ChatGPT can feel like the Hot-or-Cold guessing game, finding the right prompt to get the response you want! We examine the cognitive challenges people face in translating their intentions into prompts and make recommendations for GenAI designers
With much gratitude and excitement, I'd like to share that I will be joining Stanford (
@StanfordHAI
,
@StanfordEd
, and
@StanfordHCI
) as an Assistant Professor and HAI Junior Fellow starting Fall 2021.
Thrilled that our paper received a Best Paper Award at
#CHI2020
! The fun part about this project was digging into cog psych literature to understand about text comprehension to design the human-AI experience.
Paper:
🚨How do software teams apply Human-AI guidelines in practice? In our
#CHI2022
paper with
@im__jane
,
@eytanadar
,
@SeifertColleen
, we find out by interviewing 21 practitioners (UX, engineers, data scientists, managers) from 14 organizations.
Paper: 🧵 1/n
I am on the job market, and I am looking for faculty/ research positions. I work on HCI challenges to Applied AI (Tools for HAI design, Responsible AI, AI systems for Creativity/ Education). Please RT and reach out if you know of any position.🙏
I am looking for Ph.D. students interested in HCI+AI/ AI in Education. Please apply to
@stanfordHCI
and
@StanfordEd
and reach out if you have questions! Also, check out this awesome list by
@andrewkuznet
👇
Considering a CS/HCI PhD this year?
Once again, I've made a list of open research positions to browse by subject area! 👇
If you/peers are hiring students, I'd love to add your recruitment tweet and make sure students can discover your lab in time!
We (
@SeifertColleen
,
@eytanadar
,
@umsi
) won the Best Paper award at
#IUI2021
!!! This is my first time attending IUI and I am very much inspired by this community. Special thanks to the organizers for an excellent virtual experience :)
Really excited to share this work with
@SeifertColleen
and
@eytanadar
on Prototyping AI-Powered Interfaces. We propose a new workflow and tool (ProtoAI) that makes AI accessible to UI/UX designers.
#IUI2021
1/4
Paper:
Video:
Woke up feeling like my throat was on fire, tested positive for COVID. Will reach out to people individually, but if we had hallway chats at
#CHI2022
you might want to get tested.
Can anyone recommend recent papers that use LLMs to simulate humans ( teachers, students, mental health patients, HCI ‘users’, historical figures, etc)?
If you are a current HCI student in the US with experience in AR/VR software development and interested in accessibility looking for an internship (summer and fall) reach out to me!
📢Introducing AltCanvas- a new tool blending generative AI with a tile-based interface to empower visually impaired creators! AltCanvas combines a novel tile-based authoring approach with Generative AI, allowing users to build complex scenes piece by piece.
#assets2024
Just arrived in Honolulu for
#CHI2024
☀️, if you are around and want to chat about AI and education or contemplate the future of HCI research, do reach out!! Students coffee (or Ube latte) is on me!! ☕️
📢
#CHI2024
Paper: Using tools like ChatGPT can feel like the Hot-or-Cold guessing game, finding the right prompt to get the response you want! We examine the cognitive challenges people face in translating their intentions into prompts and make recommendations for GenAI designers
📢Current
#ResponsibleAI
research focuses on ethics post-deployment but less so on ethical issues in ML evaluation. We propose a framework to consider the tradeoffs between information gain and potential ethical harms. (with
@JessicaHullman
and Neha Gupta)
📢New Paper: Do model cards help teachers plan pedagogy? In our new
#CHI2024
paper,
@meiflwr_
looks at the information needs of teachers to effectively integrate ChatGPT into classroom instructions and student assignments. Findings from our interviews with 22 teachers:
Thank you, survivors, for the immense courage in sharing your stories! Walter and community members who still support him and harass survivors-- fuck off!
@umichCSE
, I hope you can (and will) do better!
Walter Lasecki assaulted me. He pinned me against a bar, he put his hand up my skirt and his fingers in my underwear. He grabbed and twisted. He leaned in and insisted that I sleep with him. “It could be innocent,” he said.
More often than not, the AI subcomponents are developed first and then HCI people are brought in to shape the human experience around AI. This is obviously problematic! Here is an alternative to design human-centered AI from the GROUND UP:
I made this comic last year based on a conversation with
@eytanadar
about collaborative design in HAI. Wondering if I should include it in my dissertation🤔
I am planning to teach a new course on designing explorable explanations for learning. I’d appreciate pointers to relevant courses/ reading materials you think I should include/ tools for programming novices to create interactive simulations.
Really excited to share this work with
@SeifertColleen
and
@eytanadar
on Prototyping AI-Powered Interfaces. We propose a new workflow and tool (ProtoAI) that makes AI accessible to UI/UX designers.
#IUI2021
1/4
Paper:
Video:
Check out this new work on using
#GPT
to scaffold exploratory creative coding, led by some amazing students! Our design allows artists to seamlessly switch between NL prompts, dynamic prompt-driven GUI, and direct code editing.
@StanfordEd
@StanfordHAI
@StanfordHCI
#UIST2023
Pleased to share Spellburst: an LLM-powered creative coding tool, accepted at
@ACMUIST
! Artists can move between semantic (high level) and syntactic (low level) ideas and explore many branches in parallel
Paper:
More on the
@Replit
/
@StanfordHCI
collab:
In our new
#DIS2021
paper with
@SeifertColleen
and
@eytanadar
, we explore how HCI and AI practitioners can co-create both the AI and UX (collectively AI experience or AIX) from the ground up.
Paper:
1/n
1/ 📝 New paper alert! 🚀 We take a closer look at claims made in research around VIS4ML, which presents tools that use visualization to support human involvement in ML pipelines, to make ML models human interpretable & actionable
paper:
w/
@JessicaHullman
My neighbor in India died from covid a few days ago and I now wake up 15 times at night to check my phone for bad news... How are you all coping? Appreciate advice/ ways to offer help! 🙏🏽
Really enjoyed
@HariSubramonyam
’s talk at
@cmuhcii
about using leaky abstractions and data probes to facilitate the co-creation of human-centered AI. Super cool research!
Thanks so much
@HariSubramonyam
for sharing fantastic work on AI experience design at the
@StanfordHCI
seminar! It would be exciting to explore how leaky artifacts invite more diverse stakeholders to participate in the AI development process.
Hope you are having a great
#CHI2023
day 1! Tomorrow we will present both papers on supporting designers to work with pre-trained AI models in the 11:10 session "AI Trust, Transparency and Fairness" in room Y05+06 (
@StevenM80197238
will have a much better presentation than mine!)
Based on my twitter feed CSCW feels like a three-day-long-bear-hug, everyone is so nice and supportive... Or maybe I just follow all the nice people who attend CSCW :)
🚨 Calling all science teachers! 🚨 We're conducting a research study (paid, on Zoom) to test a new generative AI tool for designing low-fidelity interactive simulations.
Interested? Fill out this form:
Please share!
#EdTech
@StanfordEd
@StanfordHAI
Flight attendant: Is there a doctor onboard?
Dad: *nudging me* that should've been you
Me: Not now Dad
Dad: Not asking for a HCI Researcher to help, are they?
Me: Dad, there's a medical emergency...
Dad: Go and see if talking about “ design implications" help
Excited to share our
#acmDIS2021
paper on a process model to bring HCI and AI practitioners together to co-create AI experiences.
#HAI
#AI
#UX
@umsi
Check out our talk :
P2: How can HCI and AI practitioners co-design both the AI and UX (collectively AI experience or AIX) from the ground up? Subramonyam et al.
@umsi
propose a collaboration process model for AIX design.
#hai
#uxdesign
#dataindesign
#DIS2021
Paper:
Are you going to
#CHI2024
? Consider signing up for our course, “Human-Computer Interaction and AI: What Practitioners Need to Know to Design and Build Effective AI Systems from a Human Perspective.”
We’re Course 07 to be held on Wednesday, May 15, 2024.
Pre-trained models are readily available to product/UX designers. But how can we help them decide which model to use and what tasks to support? Read our new papers on design-oriented documentation and tooling for model interrogation
#CHI2023
@StanfordHCI
@StanfordHAI
@StanfordEd
🧵sharing two
#CHI2023
papers!
The availability of pre-trained models (incl.
#GPT
) lowers the barrier to building on AI capabilities, but also raises questions about how to make responsible decisions on whether, where, and how to use them, and who should make these decisions
I did an interview with the wonderful Kathy Miller
@StanfordHAI
. Excited about this research direction and new projects looking at collaboration support tools for human-centered AI
Could collaborative design processes lead to better AI? We spoke with HAI Faculty Fellow
@HariSubramonyam
about his vision of bringing diverse academic communities together to build human-centered AI.
Curious about the future of Human-AI interchange? 🤔Us too! We dream of AI tools that can dynamically respond to us, adjusting their representations and conceptualizations based on context and feedback. Join us for lunch at
#CHI2024
on 5/13 to chat about
#dynamicabstractions
👇🏽
I rarely tweet, but this is work sharing: found out this morning that our paper on "Affinity Lens: Data Assisted Affinity Diagramming with Augmented Reality" with
@eytanadar
and Steven Drucker got a best paper at
#chi2019
:) Link to paper:
Check out our new article on how UX can help operationalize Responsible AI values. Thanks for leading the effort
@QVeraLiao
, it was fun collaborating with you,
@mihaela_v
, and
@lgw_
!
@StanfordHAI
Happy to see our article on the role of UX in Responsible AI out as
@interactionsMag
's cover article this month
Great fun to write this one with people whose works have been huge inspirations
@mihaela_v
@HariSubramonyam
@lgw_
I got to spend the past 10 days in Helsinki/ Aalto University (talk recording soon), but here are my top three things about Finland:
1. Authentic conversations and "strength in vulnerability"
2. Great place for an introvert (me)
3. Best free hotel breakfast!!!
I was at UIST'19 and saw Walter hand out several drink tickets to students and plan to meet late at night for practice talks. I wish I hadn't dismissed my surprise at his behavior as an advisor. I must do better. We as a community must do better!
If you are working in a multidisciplinary team to create human-AI applications, read the article (and our paper) for ways to reduce friction during collaboration.
#AIExperienceDesign
“UX designers use their understanding of human behavior and usability principles to design graphical user interfaces. But AI is changing what interfaces look like and how they operate,” says HAI Faculty Affiliate and
@StanfordEd
Professor
@HariSubramonyam
.
Especially if you’re early career, my advice for
#chi2022
would be to spend more time making friends with people in a similar point in their career to yourself (yes, making friends, not networking!) than trying to get noticed by senior ‘superstar’ folks
I will be at
#CSCW2023
from Sunday evening to Tuesday night. Eager to reconnect with old friends and make new friends in the field. If you're around, let's catch up!
Ok, I am too distracted to do any real work. So I took a little time to start my blog. Article 1: A short reflection of my Ph.D. journey so far. (spoiler alert: it has mostly been good!)
Validation for this type of behavior comes in many forms. Arrogance, snark, and silence are enough validation. For all of us starting as new faculty, I hope we will treat everyone with humility and respect and demonstrate to our advisees to do the same. We cannot remain silent!
I was able to meet up with half of the folks I started my PhD with back in 2015 and now we’re all going on our own adventures. I was lucky to have such a great cohort to grow with.
🚨How do software teams apply Human-AI guidelines in practice? In our
#CHI2022
paper with
@im__jane
,
@eytanadar
,
@SeifertColleen
, we find out by interviewing 21 practitioners (UX, engineers, data scientists, managers) from 14 organizations.
Paper: 🧵 1/n
We are exploring ways to involve multiple stakeholders in defining ML data needs. If you are a UX professional with AI design experience , ML engineer, or legal/ethics practitioner, please consider participating in our research. Please RT
@StanfordHAI
@StanfordEthics
I'm seeking a new PhD student at
@NorthwesternCS
! Interests include modeling visual analysis/learning from data under uncertainty, interface-mediated feedback loops, implications & decision making w/differential privacy. Join my lab w/
@mjskay
! Pls retweet!
My favorite slide for tomorrow's lecture on "Human-Generative AI Interfaces: Promises, pitfalls, and the Role of Design." (Weirdly, the two tweets showed up in close proximity on my Twitter timeline)
It was a pleasure hosting Nicky Case
@ncasenmare
in our class this week to talk about their process and advice for creating explorable explanations
@StanfordEd
Super excited to share our work "VideoSticker" - a tool for taking notes from videos in a semi-automated manner. This work started during my master's
@UMich
with amazing
@HariSubramonyam
@eytanadar
.
#IUI2022
📎Paper:
📽️Video:
~Three weeks into lockdown is when people start checking their
#LinkedIn
messages from six years ago. Someone I reached out to for internships back in 2014 finally got back to schedule an interview for this week :)
Generating images for a talk using Bing... The prompt: "person who is a UX designer and an engineer, cartoon, with multiple hands, some holding UX design tools others holding software programming symbols" the output:
HAI is a multidisciplinary endeavor! If, like me, your research spans several disciplines and domains, this is an excellent opportunity to advance your research agenda while collaborating with awesome people from many departments. Come join us!!!
Winners of the 2019 Microsoft Research Dissertation Grants have been selected. Find out who they are and how the funding will help advance their work in AI, systems, healthcare, and other research areas:
Does anyone have good resources on what the ML community understands as Theory of Mind? It’s used a lot lately in the context of LLMs but seems different from how psychologists define ToM.
Had a wonderful time presenting at
@ACMIUI
, Greenville, USA! In our
#IUI2024
paper, we investigated how effective LLMs are for software help-seeking and introduced SoftAIBot providing software guidance using RAG and in-context prompt guidelines.
@parmitchilana
@HariSubramonyam
Want to learn more about teachers' challenges and needs when planning classroom pedagogy with LLMS?
@meiflwr_
will present our paper on limitations of teacher-facing model cards and the need for interactive experimentation at 2:15 PM Monday | Room 316B
@StanfordEd
@StanfordHCI
📢New Paper: Do model cards help teachers plan pedagogy? In our new
#CHI2024
paper,
@meiflwr_
looks at the information needs of teachers to effectively integrate ChatGPT into classroom instructions and student assignments. Findings from our interviews with 22 teachers:
📢 CFP for the 2nd
#ieeevis
#EduVis
workshop is out!! I am personally stoked to join this fantastic team that ran a super successful workshop last year! Multiple tracks and submission formats this year, including publishing your work in the awesome Nightingale Magazine👇🏽
Check out
@jasmineyshih
LBW poster on *CellSync* -- a tool that allows domain experts to collaborate with data scientists in upstream ML data and modeling tasks
#CHI2024
afternoon coffee break at 3:20 PM!
#CHI2024
Late Breaking Work Paper: It’s challenging for domain experts and data scientists to collaborate continuously in the upstream machine learning pipeline. In our system, CellSync, we utilize LLMs to better support domain experts in collaborative predictive modeling tasks.
Fun conversation with my Lyft driver (she moved here from Ukraine and it’s her first day on the job) : 1. She is learning python and told me
@drchuck
is awesome, 2. She is also learning English and discussed challenges discovering domain specific vocabulary ( dentist visit)…
📢PAID Interview Study: If you are a UX professional or PM working on AI/ML applications, we'd love to talk to you about your design ideation process. Please sign up:
RT appreciated
@StanfordHAI
@mluxmeetup
#HCI
📢 Do you often engage in design ideation in your role as UI/UX designer, UXR, PM, or others?
Have you worked on AI/ML apps?
🌟 If so, pls consider joining our paid interview study to explore how to support design ideation with AI!
👉 Info and sign-up:
Booked tickets to India to see family but I feel I am going to have to cancel or risk being stuck because of travel bans 😬 What’s everyone else doing about their winter break (international) travel plans?
📢 Do you have a middle or high schooler interested in learning science through interactive simulations? We're seeking students to participate in a research study using a new generative AI tool! 🧪✨
Please fill out our eligibility form:
In the Bay Area, Lyft drivers are great interview participants to learn about software design challenges. I should carry a stack of IRB consent forms in my backpack!
With Sherry Wu, Mina Lee, Ken Holstein, Vera Liao, Hari Subramonyom, and Min Kyung Lee, I am organizing a CSCW panel to discuss how LLMs could influence CSCW and social computing research & vice versa.
What would you want to ask the panel?
Ok, so this is full-on shameless self promotion, but I love the way that this blog post tells the story behind the work that
@hannawallach
and I have been doing with our FATE/Aether colleagues for the past few years. It makes me feel energized all over again!