iiinova — where your digital self keeps evolving

Carolyn Baer @carebaer7.bsky.social · 9h

Meta and OpenAI released their agents Muse and Dots. They use cartoon characters as mascots, the Joe Camel of AI marketing. This anthropomorphism is 1) bad for children’s understanding of AI and 2) intentionally chosen for this reason.

111

Reposted by Adam K. Dubé

I'm recruiting a graduate student for Fall 2027! Full funding, beautiful campus, and a great community to work in. International applicants welcome - Canada is a great place to be. More info about what I look for and the app process: www.trentdevminds.ca/join-the-lab www.trentu.ca/graduatestud...

0129

We talk about these type of design decisions in our policy brief on how to regulate AI. bsky.app/profile/edte...

000

Portraying it as such is a choice, one that is meant to shape how people see AI and its relationship with us. It may seem silly to regulate, but we stopped Joe Camel from selling cigarettes. What do you think?

100

AI companies are targeting both youth and AI-resistant adults with this marketing choice. A friendly cartoon can’t be bad for the environment, take your job, be confused for a human, or end society. But, it does not have to be anthropomorphic at all. It’s not a being

100

Anna Mehler Paperny @mehlerpaperny.bsky.social · 30/09/2026

111

Reposted by Adam K. Dubé

Indigenous people, overrepresented in Canada's prisons, are far more overrepresented in a hidden form of solitary meant for suicidal and self-harming prisoners. ijb.utoronto.ca/news/hidden-...

Line chart shows Canada's use of observation cells for Indigenous and white prisoners.

33325

Truthfully, ease of collecting self-reports of parental phone use and parental reports of child behaviour and wellbeing are producing a large body of correlational data on this topic in a short period of time. The underlying theory explaining the phenomena is ill-considered.

This is not intentional of the researchers. I have spoken with people in this space at SRCD. They are not thinking of these studies in this way. It does not occur to them there is an issue in the underlying method, interpretation, and recommendations.

Yes. But not the analyses. I have a problem with underlying assumptions of the research area. It assumes that there is some ideal state of parenting that exists without phones. It ignores who the ‘parents’ are in the samples. It recommends parental best-practices that will be born by women.

Phone snubbing. It is the phrase used in this research space to describe the practice of parents attending to phones over kids.

Studies of bad parental tech habits are often just ways to blame women for not moming hard enough.

000

Phone use during dinner time, sure, let’s have a conversation. But let’s bring dads into it, too.

Mom works all day, still goes to all sports games or dance recitals. God forbid if they read the news on their phone during these hours long events. What are we doing.

100

Invokes a long debunked idea, I assume this is the journalist and not the researcher.

100

Adam K. Dubé @edtechprof.bsky.social · 29/09/2026

Not a fan of past Phubbing research It blames women for not being attentive to children Studies are of parents; moms participate, men less Assumes a Rockwell scene, sans phones, moms dotting on kids TV, radio, newspapers, books All argued to distract women from ‘motherly duties’

101

Now I will have to check it out for spooky season. I highly recommend the new Resident Evil movie. It’s great.

Adam K. Dubé @edtechprof.bsky.social · 25/09/2026

Love this thread. Not sure intensive and ambitious interventions would be more effective. Kraft (2020) studied 747 educational interventions and found avg d = .16 Studies with n > 2000, d = .05. These are targeted, designed, effortful interventions. Real change/learning is just hard.

journals.sagepub.com

Sage Journals: Discover world-class research

Subscription and open access journals from Sage, the world's leading independent academic publisher.

130

Adam K. Dubé @edtechprof.bsky.social · 25/09/2026

Could behavioral change interventions be more effective if they were more ambitious? Maybe. But, Kraft (2020) studied 747 educational interventions and found avg d = .16 Studies with n > 2000, d = .05. These are targeted, designed, effortful interventions. Change/learning is hard. That’s ok

doi.org

Sage Journals: Discover world-class research

Subscription and open access journals from Sage, the world's leading independent academic publisher.

033

Reposted by Adam K. Dubé

Nathan Holbert @snowdaylearninglab.org · 24/09/2026

Kids repurposing tech to meet their extremely unique needs leaving adults baffled is one of my absolute favorite reoccurring stories

0126

Reposted by Adam K. Dubé

Carsten Schwemmer @cschwem2er.bsky.social · 28/07/2026

1/6 What happens when you tell an AI browser: “Fill out this survey”? It fills out the survey. In our new study, led by @rtopinkova.bsky.social, we tested three consumer-facing agentic browsers without lengthy prompts, custom code, or technical fine-tuning.

33728

Adam K. Dubé @edtechprof.bsky.social · 23/09/2026

I contributed to a new national policy brief on Bill C-34, the Safe Social Media Act, submitted to the House of Commons. Our recommendations focus on protecting youth through stronger regulation of digital platform design, privacy, transparency, and developmentally appropriate safeguards.

tlclab.owlstown.net

TLC Lab - TLC Lab Contributes to National Policy Brief on Protecting

Dr. Adam Dubé, Director of the Technology, Learning, and Cognition (TLC) Lab, contributed to a new national policy brief submitted to the House of Commons St...

020

Adam K. Dubé @edtechprof.bsky.social · 15/09/2026

“Only experts should use AI, not learners.” But what if using AI deskills experts? Oncology doctors became less skilled as they used AI in their profession. A 1 year study.

031

Adam K. Dubé @edtechprof.bsky.social · 15/09/2026

@brianaltano.bsky.social Thank you for all you did at ign. I listened to every beyond and laughed at every dumb joke. I look forward to whatever you get up to next.

000

Adam K. Dubé @edtechprof.bsky.social · 09/09/2026

New Apple Watch 12 transcribes all your conversations. Think Different Think Panopticon

000

You may like this. “AGAINST THE UNCRITICAL ADOPTION OF ‘AI’ TECHNOLOGIES IN ACADEMIA” hcommons.org?get_group_do...

hcommons.org

131

20 points is about 1 year of learning loss. More frequent AI use = more loss

000

How students use AI impacts learning differently. Using it to draft assignments, worst outcome.

100

PISA data is out today. A lot of info on AI use and learning. A thread with Canadian insights Fist, higher AI use negatively related to Science learning. Canadian students use it less, perform better than overage.

220

Accuracy of models by model complexity.

000

Tobias Gerstenberg @tobigerstenberg.bsky.social · 03/09/2026

Great analysis on state of hallucination rates for AI. Gives index for AI getting things confidently wrong. Better models = more confident hallucinations. Figure shows models’ rate of confident hallucinations. Lower is better Y axis is % of incorrect outputs that are confident

130

Reposted by Adam K. Dubé

The Stanford Psychology Department is having a search for a tenure-track Assistant Professor with a research focus on developmental and/or social psychology. Application deadline is 10/15. More infos here: facultypositions.stanford.edu/jobs/assista...

facultypositions.stanford.edu

Assistant Professor, Psychology - Stanford University, California, United States

The Department of Psychology at Stanford University is seeking applicants for a tenure-track Assistant Professor position, with a research focus on developmental and/or social psychology. In addition ...

03132

Adam K. Dubé @edtechprof.bsky.social · 03/09/2026

Great chat about tech and schools!

032

Reposted by Adam K. Dubé

LEARN Pedagogical Services and Educational Technology Team @learnped.bsky.social · 03/09/2026

Dr. Adam Dubé shares the ONE question every educator needs to ask school leadership before adopting new EdTech tools. 🎧 Stream the full episode of the Shift Ed Podcast for a research-backed perspective on the future of teaching! www.buzzsprout.com/2404304/epis... #ShiftEdPodcast #AI #AIinEducation

111

Reposted by Adam K. Dubé

Steven R. Shaw, PhD @shawpsych.bsky.social · 30/08/2026

For Canadians looking to move away from using Google docs, I recommend cdox.ca. Quality small tech. Simple and easy to share with co-authors. I've used it for several months and found it to be a solid platform.

cdox.ca

cDox | A Google Docs alternative with no AI training.

A private alternative to Google Docs and Sheets. Hosted on independent bare metal servers in Canada. No AI training, no data extraction.

196

Adam K. Dubé @edtechprof.bsky.social · 30/08/2026

Also viewable in other countries.

Adam K. Dubé @edtechprof.bsky.social · 30/08/2026

Google renamed Lake Ontario lake america in the USA and some other dragons. Not in Canada. Not capitulation, cooperation.

000

Adam K. Dubé @edtechprof.bsky.social · 29/08/2026

Not that pubs = impact, but I was curious what Carl had written and it is very little. Such an outsized influence for someone who has not done much original scholarship. Maybe I’m missing something. So disheartening to see this stuff pass for what LS is.

Adam K. Dubé @edtechprof.bsky.social · 28/08/2026

Give Claude students’ TikTok and instagram. Teaching agents write personalized tasks at intersection of curriculum and market critical skills Learner agents produce pitch decks and zero-touch apps tuned on students’ social media Teaching agent verifies student agent Produce ad Infinitum

000

Adam K. Dubé @edtechprof.bsky.social · 28/08/2026

Horrible take. As a postdoc at U of Toronto, we unionized. Before, some post docs got $70k/yr, some $20k/yr. After, $50k minimum. Profs thought it was ok to pay a PhD well below poverty rates; sure, you lack funds, but you still get the labour Can’t be a mentor and a master.

0388

Adam K. Dubé @edtechprof.bsky.social · 27/08/2026

Start the watch. Will Google cave?