Hard recommend for the Mystery AI Hype Theatre 3000 Podcast - although you will end up sitting with your head in your hands, maybe crying a bit…

Imagine believing that using text-generating machines to perform clinical assessments & replace expert advice from real, qualified humans could improve health care. Such machines will deskill experts & surely kill. And you can bet the machines & their makers won’t be held accountable.

Bureaucratic benchmarks are soul-crushing because they leave a big gap between what we care about and what can be measured. When we forget about the things we actually care about (like making interesting discoveries) and we write worse papers to get more publications, the metric eats the value.

If your study is framed as asking whether "AI" does X as well as humans do, it's fundamentally misguided and I'd argue not scientifically sound. A short 🧵>>

'to treat peer review as a throughput problem is to misunderstand what is at stake. Review is not simply a production stage in the research pipeline; it is one of the few remaining spaces where the scientific community talks to itself.' 1/3

I hate this. I hate that scholars and teachers are supposed to be digital fraud experts. I hate that this part of their job description is becoming larger and larger. I hate the widening distrust. I hate a culture that aggressively devalues the curiosity and humility required for ongoing learning.

Will do a brief thread with highlights: “Many of our contemporaries now want to combine the worst of these two worlds [i.e., Psychology and Artificial Intelligence]. What could possibly go wrong? Quite a lot.” 2/🧵

It is EXHAUSTING not only being made responsible for coming up with new kinds of assignments for our students; it's also tedious reading op-eds that suggest the core problem is a crisis in teaching. But, as Chris and I lay out here, this isn't a crisis in teaching; it's an attack on learning.

The latest QRP (although it goes well beyond ‘questionable’ and straight into the realm of junk data fraud IMHO): LLM-hacking

Good luck drawing reliable conclusions from the answers that Qualtrics' AI model provides to your survey questions... bsky.app/profile/joac...

I wrote this brief talk on why “augmenting diversity” with LLMs is empirically unsubstantiable, conceptually flawed, and epistemically harmful and a nice surprise to see the organisers have made it public synthetic-data-workshop.github.io/papers/13.pdf

Delighted to that my grant proposal with Anita Eerland, Verbs and Eyewitness Testimony: A Multilab Registered Replication Report, has been funded by @NWO (Dutch Research Council) through OpenScience.nl. Excited to get started on the project I describe here. rolfzwaan.substack.com/p/memory-mis...

🚨 Now out in Psych Science 🚨 We report an adversarial collaboration (with @donandrewmoore.bsky.social) testing whether overconfidence is genuinely a trait The paper was led by Jabin Binnendyk & Sophia Li (who is fantastic and on the job market!) Free copy here: journals.sagepub.com/eprint/7JIYS...

Erasmus+ an accessible and well-supported programme.

"an ever-widening gap between those who do the work and those who administer it. And an even larger gap exists between those tasked with most of the teaching and those who do most of the budgeting." www.aaup.org/underclass-s... #Highered #PhDchat #research #teaching #academicsky

We know the drivers of research waste in academia are ⚠️Pressure to maximize papers and PhD students ⚠️Endless demands on time due to poor management ⚠️Stakeholders don't insist on robust quality systems to underpin mission critical work Solutions that don't address these are pointless.

“Berg's point is that AI doesn't merely automate tasks — it automates the very processes through which people develop their skills.”

The most precious commodity you have is your attention. You don’t have to waste it on poor-faith debates or arguments with strangers if you don’t think they’ll be productive. You can prioritize the things that matter to you and make your life richer.

I feel you, ancient Mongolian ceramic hedgehog. I feel you.

It’s widely known (and, I think, pretty uncontroversial) that learning requires effort — specifically, if you don’t have to work at getting the knowledge, it won’t stick. Even if an LLM could be trusted to give you correct information 100% of the time, it would be an inferior method of learning it.

Absolutely this…there are still many predators evading their comeuppance, including in my own field.

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

This seems bad on like 15 different fronts

New paper by @emilyspearing.bsky.social et al. out now in the Journal of Environmental Psychology Black Summer Arson: Examining the Impact of Climate Misinformation and Corrections on Reasoning doi.org/10.1016/j.je...

Evergreen.

“Many of our contemporaries now want to combine the worst of these two worlds. What could possibly go wrong? Quite a lot.”

A spot in hell? @emilymbender.bsky.social @alexhanna.bsky.social

After submitting a FOIA request UKRI, I obtained success rates by three grant call scheme and I can only say that I am disheartened by the results: - AHRC Responsive Mode 2025: 2% - ESRC New Investigator Grant 2025: 1% - ESRC Research Grant Round 2025: 1%

Can AI simulations of human research participants advance cognitive science? In @cp-trendscognsci.bsky.social, @lmesseri.bsky.social & I analyze this vision. We show how “AI Surrogates” entrench practices that limit the generalizability of cognitive science while aspiring to do the opposite. 1/

The anthropomorphic language that the AI industry uses to talk about its tech is a large part of the reason why people do not understand how it actually works.

I DON'T NEED TO SAVE TIME. I AM HERE SPECIFICALLY TO READ THE LONG DOCUMENT. STOP OPTIMIZING ME.

Yes - and you particularly don’t have to do this if they are predatory manipulators (esp. of young female researchers) and request lists of self-citations when reviewing…. Of their crappy, derivative, methodologically unsound work. Oh and when they invite themselves to give keynotes, just say no.

Ayup. No excuses for being this dumb. This study exists and is public: www.sciencealert.com/openai-has-a...

Day 2 of ##Hiddenref2025 and a reminder that people are as important as outputs.

Hard agree - this is insane; pretending it’s viable/OK is actually corrupt. What do advocates think they might ‘learn’ from it? Seeing the vagaries of human performance is part of the fun of research but importantly prompts further questions. This kind of data fabrication is just output grifting…

On National Poetry Day, the greatest poem I have ever read

The most depressing but plausible article I have ever read. Onwards into the “moronic inferno”…

A lot of psych is already conducted with online convenience samples & ppl are probably excited about silicon samples bc it would allow them to crank out more studies for even less 💸 How about we reconsider the idea that sciencey science involves collecting own data. www.science.org/content/arti...

The Network Plus for Analytical Behavioural Science for Security and Defence (NABS+) has a call for Visiting Fellowships. Part of our commitment to develop the next generation of behavioural science for security & defence through x-fertilisation of knowledge & skills. Submission deadline: 20/10/25

I wish more people got this. Science is inefficient. I'd say probably say inherently inefficient. I know we all want to hurry up and make the world a better place with those tax dollars, but I believe the harder we try to rush things, the less effective we will be in the long run.

Looks like my favorite paper on the age trajectory of happiness is finally out!!! So happy for the authors. Go check it out, it’s great.

At least clippy didn't make everyone decide learning isn't worth it anymore

Some folks confuse these things —who think the 'water movement' must be flawed, since bottled water is expensive, and contaminated water is poisonous. Clean water and #openaccess both cost $. We don't really need corporations selling packaged products, but communities funding reservoirs and taps.

Microsoft OneDrive is one of the most useless, frustrating pieces of software I have ever encountered in my life.

Glad Science collected this data (though the results are entirely unsurprising). GenAI cannot accurately summarize scientific papers, sacrificing accuracy for simplicity. And shame on publishers who are pushing genAI summaries on readers. Great way to accelerate an epistemic apocalypse.

But this is the point: those falsely accusing us of brainwashing students want to be able to do that themselves. They can’t bullshit people if they can critically evaluate information and arguments, and that’s a threat to them. Which is why universities are a primary target in the culture wars. /end

Excellent piece. We know how to improve writing ability: it's by doing more, not less of it. "LLMs do not improve one’s writing ability much like taking a taxi does not improve one’s driving ability. Students should hone their writing, thinking, and other academic skills at every opportunity."

On Prolific, "we estimate that about 34% of online study participants use LLMs to answer open-ended questions atleast some of the time..." Seems like a very timely paper for behavioural scientists using online samples: osf.io/preprints/so... ; We really need more papers on this issue

Why we need trust in science: @carlbergstrom.com nails it. This is why I’m sleuthing in retirement. Some say that talking about fraud destroys trust in science. No. Turning a blind eye to fraud and pretending it’s not happening is what destroys trust.