Joe Michaelis
@joemichaelis.bsky.social
Asst. prof at UIC. Investigating supporting social learning processes for interest and learning from a Learning Sciences and HCI perspective. Now with more robots! HRI, Social Robots, Ed Tech
https://jmich.people.uic.edu/
https://www.uiclitlab.com/
https://jmich.people.uic.edu/
https://www.uiclitlab.com/
Reposted by Joe Michaelis
🧵 Postdoc opportunity at McGill
I am recruiting a top postdoctoral researcher to join my lab at McGill through the Canada Impact+ Research Training Awards Initiative. This is a rare, well-funded opportunity to build an ambitious, multimodal EdTech research program in Canada.
I am recruiting a top postdoctoral researcher to join my lab at McGill through the Canada Impact+ Research Training Awards Initiative. This is a rare, well-funded opportunity to build an ambitious, multimodal EdTech research program in Canada.
December 20, 2025 at 1:57 PM
🧵 Postdoc opportunity at McGill
I am recruiting a top postdoctoral researcher to join my lab at McGill through the Canada Impact+ Research Training Awards Initiative. This is a rare, well-funded opportunity to build an ambitious, multimodal EdTech research program in Canada.
I am recruiting a top postdoctoral researcher to join my lab at McGill through the Canada Impact+ Research Training Awards Initiative. This is a rare, well-funded opportunity to build an ambitious, multimodal EdTech research program in Canada.
What the absolute shit is this? ACM using AI summaries instead of abstracts in search results?! So, if I understand correctly, instead of the author's own summary of the work (i.e. an abstract) we are better off with some regurgitated slop version of that summary instead? @chi.acm.org
December 16, 2025 at 8:02 PM
What the absolute shit is this? ACM using AI summaries instead of abstracts in search results?! So, if I understand correctly, instead of the author's own summary of the work (i.e. an abstract) we are better off with some regurgitated slop version of that summary instead? @chi.acm.org
Reposted by Joe Michaelis
Excited to say that I'll be running more robot comedy events this spring! ROBOT RIOT brings together roboticists, ethicists, and some of Denver's funniest comedians, to inform the public about risks of AI and Robotics while making them laugh.
Our first show is 2/6/26:
Our first show is 2/6/26:
December 12, 2025 at 2:16 PM
Excited to say that I'll be running more robot comedy events this spring! ROBOT RIOT brings together roboticists, ethicists, and some of Denver's funniest comedians, to inform the public about risks of AI and Robotics while making them laugh.
Our first show is 2/6/26:
Our first show is 2/6/26:
Get there early for www.spectaculousmusic.com
December 11, 2025 at 9:52 PM
Get there early for www.spectaculousmusic.com
For anyone who cares about rich studies that center and value children throughout the entire R&D process, this is a lovely community for your work.
IDC is the leading venue for work on child-centred design, learning, and interaction, and this year’s theme “Sustainable Futures” welcomes papers exploring fresh perspectives on how technology can support long-term, equitable futures for children. #IDC2026
idc.acm.org/2026/calls/f...
idc.acm.org/2026/calls/f...
December 9, 2025 at 6:15 PM
For anyone who cares about rich studies that center and value children throughout the entire R&D process, this is a lovely community for your work.
Reposted by Joe Michaelis
It's unboxing day for DEGREES OF FREEDOM from @mitpress.bsky.social! :D
November 23, 2025 at 5:44 PM
It's unboxing day for DEGREES OF FREEDOM from @mitpress.bsky.social! :D
Reposted by Joe Michaelis
Well, at least we didn't make online survey experiments the dominant methodological approach in major areas of the discipline
new paper by Sean Westwood:
With current technology, it is impossible to tell whether survey respondents are real or bots. Among other things, makes it easy for bad actors to manipulate outcomes. No good news here for the future of online-based survey research
With current technology, it is impossible to tell whether survey respondents are real or bots. Among other things, makes it easy for bad actors to manipulate outcomes. No good news here for the future of online-based survey research
November 18, 2025 at 7:21 PM
Well, at least we didn't make online survey experiments the dominant methodological approach in major areas of the discipline
OMG. THIS. PAPER. IS. PERFECT!!! I am going to staple the How to Assess Qualitative Research section to a boxing glove and add that as supporting material for my next HRI submission. doi.org/10.1080/0737...
November 13, 2025 at 4:54 PM
OMG. THIS. PAPER. IS. PERFECT!!! I am going to staple the How to Assess Qualitative Research section to a boxing glove and add that as supporting material for my next HRI submission. doi.org/10.1080/0737...
Reviews that don’t appropriately evaluate based on the method have been killing me lately. If I had a dollar for every review of a qual paper that criticized a lack of generalizability…
Really unfortunate that a student led project got rejected for AERA b/c one of the reviewers didn’t understand what a design case is. Their feedback was that the lack of empirical data made it a poor proposal - when that’s not a design case feature or requirement.
November 9, 2025 at 7:19 PM
Reviews that don’t appropriately evaluate based on the method have been killing me lately. If I had a dollar for every review of a qual paper that criticized a lack of generalizability…
Holy crap. I thought my conference homes were broken, but the guardrail for oversubmitting set to 25… wow
This is really unfair to my 5 phd students, each of whom were planning on submitting 6 papers with me as co-author. *sarcasm*
📢🚨Folks, you can only submit 25 papers this year to CVPR (25 submissions per author cap). So, please plan accordingly.
November 4, 2025 at 2:53 PM
Holy crap. I thought my conference homes were broken, but the guardrail for oversubmitting set to 25… wow
At the "listening to Minor Threat while grant writing" stage of disillusionment, burnout and frustration with the crumbling systems of academia. Turns out teen angst and existential threat are on the same spectrum of anxiety and rage. minorthreat.bandcamp.com/album/salad-...
Salad Days, by Minor Threat
3 track album
minorthreat.bandcamp.com
October 28, 2025 at 3:57 PM
At the "listening to Minor Threat while grant writing" stage of disillusionment, burnout and frustration with the crumbling systems of academia. Turns out teen angst and existential threat are on the same spectrum of anxiety and rage. minorthreat.bandcamp.com/album/salad-...
Reposted by Joe Michaelis
Loving this talk from @joemichaelis.bsky.social at Northwestern today—Mental models at the intersection of Human-Robot Interaction!
October 23, 2025 at 6:35 PM
Loving this talk from @joemichaelis.bsky.social at Northwestern today—Mental models at the intersection of Human-Robot Interaction!
Current status
I can't be the only one who goes down super deep literature rabbit holes when preparing talks. Only for it to result in like, a single throwaway bullet point on a backup slide.
October 17, 2025 at 4:46 PM
Current status
Reposted by Joe Michaelis
EPFL, ETH Zurich & CSCS just released Apertus, Switzerland’s first fully open-source large language model.
Trained on 15T tokens in 1,000+ languages, it’s built for transparency, responsibility & the public good.
Read more: actu.epfl.ch/news/apertus...
Trained on 15T tokens in 1,000+ languages, it’s built for transparency, responsibility & the public good.
Read more: actu.epfl.ch/news/apertus...
September 2, 2025 at 11:48 AM
EPFL, ETH Zurich & CSCS just released Apertus, Switzerland’s first fully open-source large language model.
Trained on 15T tokens in 1,000+ languages, it’s built for transparency, responsibility & the public good.
Read more: actu.epfl.ch/news/apertus...
Trained on 15T tokens in 1,000+ languages, it’s built for transparency, responsibility & the public good.
Read more: actu.epfl.ch/news/apertus...
Reposted by Joe Michaelis
We've recently concluded a classroom study in Japan on our "Mindful Math" Tutor, an LLM-embedded ITS that uses mindful language and messages to guide students learn algebra while calming down their math anxiety. If interested in testing/using/collaborating let me know! tomonag.org/mindfulmath/
October 9, 2025 at 12:06 PM
We've recently concluded a classroom study in Japan on our "Mindful Math" Tutor, an LLM-embedded ITS that uses mindful language and messages to guide students learn algebra while calming down their math anxiety. If interested in testing/using/collaborating let me know! tomonag.org/mindfulmath/
@hri-conference.bsky.social reviewer assignments started... I like to dig in to find reviewers who really can connect with a paper. First paper I found a group of authors writing on a very similar application... perfect, right!? So similar in fact, I think it may actually just be the same people🫠
October 3, 2025 at 2:10 PM
@hri-conference.bsky.social reviewer assignments started... I like to dig in to find reviewers who really can connect with a paper. First paper I found a group of authors writing on a very similar application... perfect, right!? So similar in fact, I think it may actually just be the same people🫠
Reposted by Joe Michaelis
🤖 Robots and reading anxiety
A study with 8–11 year olds found kids were less anxious reading aloud to a robot than to a person. Robots may help children build literacy skills with less stress.
🔗 doi.org/10.1126/scir...
#SciComm 🧪 #Education #Robots #Literacy
A study with 8–11 year olds found kids were less anxious reading aloud to a robot than to a person. Robots may help children build literacy skills with less stress.
🔗 doi.org/10.1126/scir...
#SciComm 🧪 #Education #Robots #Literacy
Robotic reading companions can mitigate oral reading anxiety in children
Children exhibit fewer physiological indicators of anxiety when reading aloud to a robot compared with reading to a person.
doi.org
September 16, 2025 at 8:24 AM
🤖 Robots and reading anxiety
A study with 8–11 year olds found kids were less anxious reading aloud to a robot than to a person. Robots may help children build literacy skills with less stress.
🔗 doi.org/10.1126/scir...
#SciComm 🧪 #Education #Robots #Literacy
A study with 8–11 year olds found kids were less anxious reading aloud to a robot than to a person. Robots may help children build literacy skills with less stress.
🔗 doi.org/10.1126/scir...
#SciComm 🧪 #Education #Robots #Literacy
Well my instinct to want to be included may just swallow me whole this year. What does one do if you are already an AC for @chi.acm.org and @hri-conference.bsky.social invites you to be an AC there, too? You kiss all other responsibilities goodbye and embrace the chaos, right? #HCI #HRI
September 25, 2025 at 2:35 PM
Well my instinct to want to be included may just swallow me whole this year. What does one do if you are already an AC for @chi.acm.org and @hri-conference.bsky.social invites you to be an AC there, too? You kiss all other responsibilities goodbye and embrace the chaos, right? #HCI #HRI
Reposted by Joe Michaelis
Shell shocked at how ubiquitous turning in LLM writing is. Knew it would be a lot, but literally every paper (1-2 pages) I read today was fully authored by an LLM. I thought if I made it clear I just wanted ideas, thoughts, feeling, and didn't care about grammar or structure, it would help, but nope
September 17, 2025 at 7:33 PM
Shell shocked at how ubiquitous turning in LLM writing is. Knew it would be a lot, but literally every paper (1-2 pages) I read today was fully authored by an LLM. I thought if I made it clear I just wanted ideas, thoughts, feeling, and didn't care about grammar or structure, it would help, but nope
Reposted by Joe Michaelis
🚀 Exciting news: I’m accepting new graduate students in the McGill Faculty of Education Learning Sciences! If you’re passionate about educational technology, games, AI, and learning sciences, this is your chance to join a dynamic and interdisciplinary research program.
September 16, 2025 at 2:01 AM
🚀 Exciting news: I’m accepting new graduate students in the McGill Faculty of Education Learning Sciences! If you’re passionate about educational technology, games, AI, and learning sciences, this is your chance to join a dynamic and interdisciplinary research program.
Reposted by Joe Michaelis
Psychologists, apply! Apparently we've mainly gotten CS applicants so far. I'd (selfishly) love to have a psychologist. Interests in AI in relation to e.g., implicit bias; cultural influences; fairness; perception; learning & memory; decision making etc. would be a good fit!
Brandeis is hiring for this unique interdisciplinary 2-year postdoc/lecturer position in AI and Society. 1 course/semester + collaborative research and mentoring. Great prep for TT positions! Deadline 10/1. Details in the ad. #PsychJobs #HigherEdJobs
www.higheredjobs.com/faculty/deta...
www.higheredjobs.com/faculty/deta...
Florence Levy Kay Postdoctoral Fellowship in Artificial Intelligence and Society (2025/10/01 11:59PM) Apply - HigherEdJobs
Jobs in higher education. Faculty and administrative positions at colleges and universities. Updated daily. Free to job seekers.
www.higheredjobs.com
September 15, 2025 at 8:45 PM
Psychologists, apply! Apparently we've mainly gotten CS applicants so far. I'd (selfishly) love to have a psychologist. Interests in AI in relation to e.g., implicit bias; cultural influences; fairness; perception; learning & memory; decision making etc. would be a good fit!
Just cited this. We're finding that expert teachers are telling us they *want* automation, but end up using LLM-generated content infrequently when given the option. Seems that expert fluency is pretty powerful. This METR study really helps develop that idea to provide some nuanced context.
We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers.
The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.
The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.
September 10, 2025 at 9:03 PM
Just cited this. We're finding that expert teachers are telling us they *want* automation, but end up using LLM-generated content infrequently when given the option. Seems that expert fluency is pretty powerful. This METR study really helps develop that idea to provide some nuanced context.
Two fresh articles out in Science Robotics www.science.org/journal/scir..., and on the website banner! First is a focus article "How can educational robots enhance family life? Through careful integration." We look at what it will take for in-home robots to actually be adopted by families. #HRI #HCI
September 10, 2025 at 8:04 PM
Two fresh articles out in Science Robotics www.science.org/journal/scir..., and on the website banner! First is a focus article "How can educational robots enhance family life? Through careful integration." We look at what it will take for in-home robots to actually be adopted by families. #HRI #HCI