Sign in

Richard Huskey

@richardhuskey.bsky.social
2.3K followers 1.8K following 317 posts

Associate Professor, UC Davis. PI, Cognitive Communication Science Lab. Associate Editor, Journal of Communication. Ski bum at heart. cogcommscience.com

PostsRepliesMedia
Reposted by Richard Huskey
Neil Sehgal @nsehgal.bsky.social · 08/10/2026
🚨 New paper 🚨 Are your participants using AI to take your study? We conducted the first study directly observing AI assistance in online research, spanning 700K+ Prolific submissions. We found very rare, concentrated AI use, but gaps in safeguards. arxiv.org/abs/2610.09279
13620
Richard Huskey @richardhuskey.bsky.social · 10/10/2026
cc @talha-ozudogru.bsky.social @cstrauch.bsky.social @achetverikov.bsky.social
020
Reposted by Richard Huskey
aaron bornstein @aaronbornstein.bsky.social · 07/10/2026
"Over the last 15 years, the number of managers ... has increased by 64%. Senior professionals, an employee group that includes senior administrative analysts, information technology professionals, coaches and student health physicians, has more than doubled in that time."
sacbee.com
UC is hitting its enrollment targets. But permanent faculty hiring lags
Faculty say the University of California’s reputation for academic excellence and its contribution to new ideas and innovations is at stake.
041
Reposted by Richard Huskey
Shahab Bakhtiari @shahabbakht.bsky.social · 03/10/2026
I had a feeling recently that seems to be corroborated by chats with colleagues, so I decided to share it here. A lot of systems neuroscientists driven by algorithmic-level questions are going through what looks and feels like a research depression. 🧵 (1/n) 🧠📈 #NeuroAI
139223
Reposted by Richard Huskey
Matti Vuorre @matti.vuorre.com · 03/10/2026
12210
Richard Huskey @richardhuskey.bsky.social · 03/10/2026
I’ve sat in a few webinars from them and my sense is that they are being very proactive and taking the issue very seriously, and implementing new tools (e.g. www.prolific.com/resources/in...). Although, I think this is an arms race with no end in sight
prolific.com
New authenticity checks detect AI misuse in research | Prolific
Introducing authenticity checks: Ensure genuine human responses in the age of AI
020
Reposted by Richard Huskey
Christoph Strauch @cstrauch.bsky.social · 02/10/2026
How much of a risk are LLM-generated bots for online behavioral experiments? We present results of an entirely natural language prompt-generated agent, capable of absolving any of the 26 experiments we tested it on. Most tasks are passed with data that can pass as human osf.io/preprints/ps... (1/)
59344
Reposted by Richard Huskey
Andrey Chetverikov @achetverikov.bsky.social · 02/10/2026
Some months ago, I wrote a skeptical reply to a commentary on bots in online studies in PNAS, arguing that no evidence yet exists for such bots. Together with the authors of the original commentary, we now show that bots can pass (almost?) all online cognitive studies with human-like performance 1/n
25129
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
Last thing. Check out the great threads below. They provide more details and different perspectives on the project. 🧵from @cstrauch.bsky.social bsky.app/profile/cstr... 🧵from @achetverikov.bsky.social bsky.app/profile/ache...
020
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
Many thanks to @talha-ozudogru.bsky.social for leading the project, and @achetverikov.bsky.social, Leendert Van Maanen, and @cstrauch.bsky.social for the collegial (despite adversarial) collaboration.
120
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
Online behavioral experiments make our science more robust and representative. We can't give that up. To win this arms race, we must not only study what agents can do, but how we can detect them.
141
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
The upshot? Much like was already the case for surveys, agents can now complete online behavioral experiments. My hunch is that tasks with considerable training data will be harder to detect compared to novel tasks that require behavioral agents to act as synthetic humans. We're testing that, now.
120
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
Our project included an adversarial collaboration. @achetverikov.bsky.social and I gave AGEA seven tasks it had never seen before. AGEA completed them all, although not always on the first try, or with data that looked entirely human (see the supplemental file in the preprint).
Experimental paradigms and comparative results for selected tasks. (A) Task list with hosting-platform icons: GitHub Pages, Gorilla (https://gorilla.sc), Hanover (https://psychology.hanover.edu/web.php), Local (personal computer), Pavlovia (https://pavlovia.org), JATOS (https://jatos.org), and Self-hosted (https://experiments.andreychetverikov.org). Green names indicate tasks or different implementations of existing tasks added as part of the adversarial cross-validation step. (B) N-back mean reaction time on correct trials by memory load: AGEA (blue circles) versus Harvey et al. (2005) controls (orange triangles). (C) N-back accuracy by load: AGEA (blue bars) versus Harvey’s controls (orange bars). (D) Tower of London mean solution time on solved problems by minimum moves required: AGEA (blue circles) versus Eisenberg et al. (2019) (orange triangles). (E) Tower of London accuracy (% solved) by minimum moves required: AGEA (blue bars) versus
Eisenberg et al. (orange bars). (F) Wisconsin Card Sorting Test (64 cards): total and perseverative errors for AGEA (open blue circles) versus Marquine et al. (2021; orange): line = reported range, band = mean ±1.96 standard deviations (SD) bounded by that range, tick = mean. Notes: In B and D, orange error bars and shaded bands show reference means ±1.96 SD. In C and E, error bars on the orange bars show reference mean accuracy ±1.96 SD, bounded to 0–100%; numerical labels indicate percentages. SD reflects variation across participants’ individual scores within each condition; these intervals describe between-participant variability, not uncertainty in the mean. AGEA values come from one session per task and have no error bars.
120
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
.@talha-ozudogru.bsky.social built AGEA, a domain-general behavioral experiment agent, using off-the shelf LLM tools and prompting alone. AGEA completed all behavioral tasks we tested it on.
120
Richard Huskey @richardhuskey.bsky.social · 02/10/2026
🧵In April, I built an AI agent that could complete the Attention Network Task. The agent was task-specific and costly. I concluded that online behavioral experiments were safe from agentic infiltration. Six months later, my view has changed. Our new preprint explains why osf.io/preprints/ps...
osf.io
Generalizable experiment agents threaten online behavioral experiments
Online behavioral experiments rely on the assumption that participants are human. Advances in large language models (LLMs) challenge this assumption, but previous demonstrations have largely focused o...
2155
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
Want to learn more? Check out this great @ucdlands.bsky.social piece by Alex Russell lettersandsciencemag.ucdavis.edu/media-choices-mental-health Article: doi.org/10.1007/s44382-026-00040-5 Big THANK YOU to @gongxuanjun.bsky.social & @aeden.bsky.social this wouldn't have been possible without you!
lettersandsciencemag.ucdavis.edu
How to Hack Media Choices to Improve Mental Health, According to New Research
New research in communication finds that college students who are lonely tend to choose media likely to make them feel worse while those who are depressed or anxious tend to choose media that might im...
054
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
So. What is a public research university for? For students like Valerie, and the people her research might help.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
Valerie arrived at UC Davis with a question about her peers. She learned the skills necessary to turn that question into a research project that is now part of the published scientific record. And she learned how science can help people make better choices in their everyday lives.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
Knowing how a choice will affect us helps us make better choices; ones that resist algorithmic recommendations that are focused on maximizing engagement and not necessarily our mental health.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
This matters because, even in an algorithmically curated world, we still choose what to watch. And those choices can shape how we feel, either in ways that help us cope, or in ways that make things worse.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
Watching a calm movie is no cure for depression or anxiety, but it may temporarily make someone feel better. But negative movies might make lonely people feel worse.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
But things get interesting if you examine the role of mental health. Depressed and anxious participants chose calmer movies. Surprisingly, lonely people chose more negative movies.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
More than 300 undergraduates at the @ucdavis.bsky.social made hundreds of choices about movies they might watch. Some movies were upbeat, others grim. Some were relaxing, whereas others were intense. Valerie found that participants showed a preference for grim and intense movies.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
Her question was urgent. A national survey of nearly 85,000 students showed that 37% experience moderate-to-severe depression, 33% report moderate-to-severe anxiety, and 58% are lonely. Valerie wanted to know if students choose media that might help, or harm, their mental health.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
Valerie joined my lab as a freshman and worked in the lab all four years of her undergraduate training. She was focused on the mental health of her fellow undergraduates. She connected that to the lab's expertise on media and decision making.
120
Richard Huskey @richardhuskey.bsky.social · 29/09/2026
🧵 What is a public research university for? There's no single answer! It conducts research that helps us understand ourselves and the world around us. It trains the next generation of leaders and innovators. Valerie Klein shows what happens when these things intersect.
Valerie Klein presents here research at UC Davis
2137
Reposted by Richard Huskey
Rob Chavez @robchavez.bsky.social · 29/09/2026
Our latest paper led by Taylor Guthrie with former undergrad RA Aussie Frost. We show that multivariate fMRI signals can decode the identity of specific individuals within the DMN and in the absence of visual features or other lower-level sensory information. link.springer.com/article/10.3...
12312
Reposted by Richard Huskey
Amitai Shenhav @ashenhav.bsky.social · 28/09/2026
2 yrs ago, I briefly laid out a new theory of how motivation works (AGH). Here, I expand on the problems AGH was trying to solve, why they are hard to solve, what my solution looks like, and how this should change our understanding of thoughts & actions. osf.io/preprints/ps... 🧵A few highlights:
17542
Richard Huskey @richardhuskey.bsky.social · 22/09/2026
Also, our original thread that explains the preprint @dougaparry.bsky.social mentioned: bsky.app/profile/rich...
010
Richard Huskey @richardhuskey.bsky.social · 22/09/2026
One of our current solutions (and we’re not the only lab to be doing this) is to run a parallel in-person data collection with the ambition of having human ground-truth data against which we can train a classifier. With rapid advances in LLMs, I’m unsure how successful we’ll be in this arms race
141
Richard Huskey @richardhuskey.bsky.social · 22/09/2026
Yep @micahgallen.com, a few of us have been working in this space. Relevant threads below, if it helps. Efforts from @cstrauch.bsky.social & colleagues bsky.app/profile/cstr... And… bsky.app/profile/cstr... Comments from @achetverikov.bsky.social bsky.app/profile/ache...
achetverikov.bsky.social
Andrey Chetverikov (@achetverikov.bsky.social)
Associate Professor in Cognitive Psychology at the University of Bergen, Norway. I study decision-making and biases in perception and visual working memory, with occasional forays into higher level de...
250
Reposted by Richard Huskey
Micah G. Allen @micahgallen.com · 22/09/2026
So… we just found out that GPT sol can complete a complex online behavioral/cognitive study producing nearly indistinguishable behavior and subjective reports… right as we are about to launch. Is this doomsday for online cognitive testing?
3212669
Reposted by Richard Huskey
Jordan Deakin @jordantdeakin.bsky.social · 17/09/2026
📣 NEW PREPRINT ALERT! 📣 📝Preprint: www.biorxiv.org/content/10.6... When choosing between options with multiple attributes, what does the brain actually compare - individual attributes or integrated option values? We combined EEG + eye-tracking to find out... and found evidence for the latter 🧠
biorxiv.org
Fixation-evoked potentials reveal neural signatures of hierarchical value-integration during decision-making
Many decisions require maintaining beliefs across multiple levels of representation, from sampling and integrating information to forming action plans. To elucidate these processes and reveal the neur...
14419
Reposted by Richard Huskey
Mariam Aly @mariamaly.bsky.social · 15/09/2026
How do keep a narrative active in our mind despite interruptions? The default mode network can maintain a representation of an interrupted story over a delay, even when the delay is filled with a demanding task. This representation is strikingly transformed from the initial experience!
biorxiv.org
Cortical and Hippocampal Pathways to Mental Continuity
How do people preserve mental continuity when tasks are interrupted? One established pathway runs through the hippocampus, which stores snapshots of our mental context for later reinstatement. We prop...
17824
Richard Huskey @richardhuskey.bsky.social · 16/09/2026
Yossarian was moved very deeply by the absolute simplicity of this clause of Catch-22 and let out a respectful whistle.
010
Richard Huskey @richardhuskey.bsky.social · 16/09/2026
Latest university guidance on phishing emails
A fictitious page from the real book Catch-22. Test says: 

There was only one catch and that was Catch-22, which specified that an email from a suspicious account was not to be opened because it might be a phishing attack. But phishing attacks were often sent from legitimate accounts that had been compromised, so an email from a legitimate account could not be opened either, because it might be a phishing attack. Faculty would be crazy to open emails from suspicious accounts and sane if they didn’t, but if they were sane, they would have to avoid suspicious emails. If they avoided them, they might miss a legitimate message; but if they opened them, they might be phished. And if they received an email from someone they knew, they couldn’t safely open that either, because their account might have been compromised.
172
Richard Huskey @richardhuskey.bsky.social · 15/09/2026
Don’t miss Rachael Kee at #csf26! She’s presenting work laying out the theoretical foundation for high-throughput communication science. Then she demos Inoxity, an open-source iOS app she’s pioneered to measure temporally rich, multimodal, self-report, behavioral, biological, and digital trace data
151
Reposted by Richard Huskey
Mark Thornton @markthornton.bsky.social · 10/09/2026
Come join @elisabaek.bsky.social & me on the SANS social media committee!
021
Reposted by Richard Huskey
Mark Thornton @markthornton.bsky.social · 08/09/2026
Excited to share this new preprint from @lindseytepfer.bsky.social, @chujunlin.bsky.social, and me, in which we investigate the different ways our brains allow us to change our minds about people! "The neurocomputational mechanisms of naturalistic trait impression updating" osf.io/preprints/ps...
Schematics of scanner and online experimental designs. A) Scanner participants first completed individual difference surveys regarding their demographics, attitudes towards social groups, mood, and personality. In the MRI scanner, they then watched short autobiographical narrative videos. After the scan, they made ratings about the traits of the individuals in the videos they viewed. B) Online participants first made ratings about the traits of the individuals in the videos based on still frame images. They then made continuous ratings of the traits of these individuals over the course of the videos.Analysis schematic. A) Spatiotemporal representational similarity analysis was used to relate patterns of brain activity over the course of SEND videos to i) continuous ratings of the traits of the individuals in those videos, ii) multimodal annotations (face, voice, and speech semantics), and iii) representations learned by neural networks to support computational operations (compression, multimodal fusion, temporal integration, and prediction). B) Intersubject representational similarity analysis was used to relate intersubject correlations in brain activity to individual differences in trait impression updating, demographics, social group attitudes, mood, and personality. Computational operations of trait impression updating. The left column illustrates brain regions within which pattern similarity is uniquely associated with representations learned by neural networks trained to engage in the corresponding computational operation: compression, multimodal fusion, temporal integration, or prediction. The right column illustrates where these effects uniquely statistically mediate the association between brain activity and trait ratings. All results control for the raw annotations of face, voice quality, and speech semantics. Results are significant (p < .05) corrected for multiple comparisons across parcels via maximal statistic permutation testing.Individual differences in trait impression updating. IS-RSA results reflecting where individual differences were associated with ISC while watching SEND videos. A) Reflects where idiosyncratic trait impression updating was correlated with ISC. B) Likewise illustrates where four categories of potentially explanatory individual differences – social group attitudes, participant demographics, participant personality (Big 5), and participant mood (PANAS) predict ISC during SEND video viewing. C) Shows where social group attitudes statistically mediate the association between idiosyncratic trait impressions and ISC. Results are significant (p < .05) corrected for multiple comparisons across parcels via maximal statistic permutation testing.
14814
Richard Huskey @richardhuskey.bsky.social · 07/09/2026
The social sciences exist to help us understand ourselves and how we navigate the world. Hard-won, albeit imperfect (humans are complicated!!), insights offer evidence-based guidance. Amidst rapid change, it is time to double-down on our ability to understand www.nytimes.com/2026/09/06/o...
nytimes.com
Opinion | We Can’t Know Our A.I. Future if We Don’t Study It (Gift Article)
Just as A.I. is poised to change the world, we’re losing our best ways of studying what that change will look like.
040
Reposted by Richard Huskey
aaron bornstein @aaronbornstein.bsky.social · 03/09/2026
UC Irvine Cognitive Sciences is hiring! Focus on social cognition, broadly construed. Come be my colleague :) I am not on the search committee but happy to talk about the campus, environment, careers, etc any time. recruit.ap.uci.edu/JPF10411
recruit.ap.uci.edu
Assistant Professor - Cognitive Sciences
University of California, Irvine is hiring. Apply now!
15963
Reposted by Richard Huskey
Emily Riederer @emilyriederer.bsky.social · 03/09/2026
For no real reason, I was looking up something about the ancestry of causal DAGs in path diagrams and came across this picture. I got to see it so you do to. HT to Sewall Wright who both set the course for a very useful analytical tool but also took the time to include cute drawings of guinea pigs
Sketch of Wright's path diagram showing the influence of heredity and environment on the inheritance of color of guinea pigs. Drawing is mostly a directed acyclic graph but includes four cute sideways drawings of guinea pigs.
817942
Reposted by Richard Huskey
Eliza Bliss-Moreau @eblissmoreau.bsky.social · 31/08/2026
My dept/area is hiring! We're open to animal species/model and research area. UCD is an amazing place to do nonhuman animal work - major 🐒 center, a band new rodent vivarium in the Psych building, and capacity to house all sorts of other animals! recruit.ucdavis.edu/JPF07812
recruit.ucdavis.edu
Assistant Professor of Psychology - Biological Bases of Behavior
University of California, Davis is hiring. Apply now!
05642
Reposted by Richard Huskey
Eliza Bliss-Moreau @eblissmoreau.bsky.social · 31/08/2026
Oof…
042
Reposted by Richard Huskey
Rachael Kee @rachaelkee.bsky.social · 20/08/2026
📖 Neurobiology of Flow and Optimal Performance, Chapter 3 in Performance Neuroscience 👉 Read the chapter here: link.springer.com/chapter/10.1...
link.springer.com
Neurobiology of Flow and Optimal Performance
Flow is an experience that occurs when an individual is completely absorbed in a task, balanced by both high task challenge and individual ability at the task. Traditionally, flow has been thought to ...
021
Reposted by Richard Huskey
Rachael Kee @rachaelkee.bsky.social · 20/08/2026
🚨 New Book Chapter Alert 🚨 Have you ever been so immersed in something that everything else seemed to disappear? An athlete “in the zone.” 🏃‍♀️
A musician completely absorbed in a performance. 🎵
Or even a researcher who looks up from their work and realizes hours have passed. 🧠
131
Reposted by Richard Huskey
Katy Milkman @katymilkman.bsky.social · 10/08/2026
A little news… here’s a preview of the new season of the #Choiceology podcast launching in two weeks 🎉 That’s right… we’ll be back in your feed with fresh episodes before you know it! podcasts.apple.com/us/podcast/c...
podcasts.apple.com
Coming Soon: A New Season of Choiceology
Podcast Episode · Choiceology with Katy Milkman · S18 Trailer · 2m
041
Richard Huskey @richardhuskey.bsky.social · 10/08/2026
11/11 Now it's your turn to follow your curiosity. For the high-level story from Alex Russell @ucdlands.bsky.social, see: lettersandsciencemag.ucdavis.edu/amazon-curio... For the technical details, the @pnasnexus.org paper is open access: doi.org/10.1093/pnas...
lettersandsciencemag.ucdavis.edu
Curiosity Makes Some Amazon Book Buyers More Willing to Risk it All for Reading Pleasure
A new study in communication scoured the book buying histories and ratings of more than 35,000 Amazon shoppers and found that while some buy books that are similar to others they have enjoyed, people ...
032
Richard Huskey @richardhuskey.bsky.social · 10/08/2026
10/n The next time you choose what to read, follow your curiosity. Let your own internal algorithm guide you to something new. You might enjoy that choice the most.
110
Richard Huskey @richardhuskey.bsky.social · 10/08/2026
9/n It feels like we are bombarded by algorithms. That our own choices are slipping away. But in fact, our minds are also running algorithms; ones that help us make choices that are aligned with our own needs and desires.
110