pouring one out for the systems thinkers who can see the abyss and have to live with the darkness

Palantir’s stated mission is to defend America. Peter Thiel is a radical libertarian who does not believe ‘freedom & democracy are compatible’. Yet, the UK is overwhelmingly dependent on Palantir’s services. As a national security threat, this is off the charts. @thenerve.news 2/

NEW: Our investigation for @thenerve_news reveals not only that Peter Thiel’s Palantir is completely enmeshed in our critical national infrastructure. It’s also the ‘cloud support’ for our nuclear weapons system 🤯 1/ www.thenerve.news/p/palantir-t...

This is worth paying attention to particularly for those who have concerns about Palantir's involvement in NHS data.... www.bmj.com/content/392/...

There is still time to register for this event on school absences and crime!

The 2nd* paper from my PhD has been published as an open access Brief Report! :) Topics: SEND, exclusion, exploitation, missing children, machine learning ("AI") Paper: doi.org/10.1080/3067... The thread below attempts to explain what its about in a way that's accessible *first solo author paper

“GPS’s remarkable accuracy rests on years of meticulous mathematical labor done behind the scenes by scientists like West, who pursued the work not for recognition, but because it mattered” ❤️

"Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them." ― Frank Herbert, Dune cassidoo.co/post/good-br...

“It’s much more about ‘how can we support you to meet the expectation’” “a “welcome hub” in which pupils can collect uniform they may need, eat breakfast and receive support from pastoral staff […] open every morning […] partnered with a charity to give parents, carers and children toiletries“

This analysis of papers from 1980 to 2025 goes beyond LLMs, but notes one challenge that arises when research questions rely on the same tools and datasets: “the adoption of AI seems to induce authors to converge on the same solutions to known problems rather than create new ones.”

"The hardest single part of building a software system is deciding precisely what to build. [...] No other part of the work so cripples the resulting system if done wrong. No other part is more difficult to rectify later." Fred Brooks, "No Silver Bullet"

An invaluable new report. A "premortem" is exactly what we need.

That's not to say LLMs are useless for legal research. You've just got to treat them like a brilliant and extremely fast trainee who is also a psychopath and pathological liar.

'The Effect of Counting Rules on Cross-National Comparisons of Homicide', by @davidbuil.bsky.social (who came up with the idea & did all of the data analysis and most of the write up), myself & @marcelo-f-aebi.bsky.social, just out at Social Indicators Research (link.springer.com/article/10.1...).

“Maybe not everything has to be digital or digitized, maybe the internet shouldn’t be inserted into absolutely everything, maybe we shouldn’t be constantly connected in the way we’re now expected to be, and generative AI certainly does not need to be forced into every facet of society.” ❤️

for all the talk about this being the next transformative technology, there is an awful lot of forcing from above. no one ever met with a country's PM to tell them smart phone adoption was too low. what are we doing here?

Our new journal @evidencebasejnl.bsky.social will officially be launching its inaugural issue in April. Evidence assessments, cutting edge methodology papers, research commentaries .. www.tandfonline.com/journals/reb...

📢 New publication – and a great way to end the year In this study, we use MAIHDA (Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy) to study who is most at risk of violence victimisation in England and Wales. Paper: doi.org/10.1177/1748... 1/n

Our work begins with a simple, but bold idea: powerful people using AI to decide our fates is not inevitable. At TechTonic Justice, we’re working for a better today and a better tomorrow. And, we want you to join us!

Big Tech is rapidly building out data centers to power AI—without adequate oversight, transparency, or accountability. @publiccitizen.bsky.social has released a new report revealing who truly benefits from the acceleration of AI across all sectors of society, and who is left to pay the price.⬇️

"Nobody is using this responsibly": @kevindeliban.bsky.social's warnings on A.I. and the lessons he took from beating the state of Arkansas in court after it denied benefits to Medicaid recipients because of decisions made by an algorithm.

Someone has messaged me on here but I can’t read it without providing data to KWS which I’m reluctant to do. If you have my contact details please email/text me. If not, perhaps reply to this post.

📢 We’re hiring a Lecturer in Criminology & Criminal Justice 📢 Please do share widely 1/3

Just a university web page… aimed at PhD students… seeking guidance on thesis submission… asking them to first ask “AI” for answers before contacting support staff, followed by a statement that the information provided by the “AI” tool can be incorrect 😧 Higher education 🤔

Closing out my year with a journal editor shocker 🧵 Checking new manuscripts today I reviewed a paper attributing 2 papers to me I did not write. A daft thing for an author to do of course. But intrigued I web searched up one of the titles and that's when it got real weird...

Academics and technologists are sounding the alarm about a growing crisis in scholarship as we know it: AI-generated citations of nonexistent papers that have infested real journals. Despite being fake, the sources are widely assumed to be authentic the more they appear in published literature.

"When the grades were released, for most of the modules they got better grades and were laughing at me.” "I am not proud of this degree anymore." 😔 “Would you want to travel over a bridge that's been designed by one of these students?" www.bbc.co.uk/news/article...

#AcademicSky Has anyone else experienced these errors on the Taylor & Francis submission portal or does anyone know how to resolve it? Seems it’s happened for others too (www.reddit.com/r/academia/c...)

It’s good to see papers start to address LLMs as structural plagiarism — provenance, more hidden than the original words or training data. www.nature.com/articles/s42...

“a source of trauma for a significant minority of vulnerable pupils”

Astonishing summary of correspondence chscp.org.uk/wp-content/u...

This practice, for example, is a form of abuse, particularly when used on children from chaotic, neglectful and/or abusive homes. A newly qualified teacher should be aware of this. So the CEO named here being appointed Chief OFSTED inspector was frankly gobsmacking. @bphillipsonmp.bsky.social

@bphillipsonmp.bsky.social needs to join the dots on this one. We have the most significant, deepening youth mental health crisis since records began. Of course children need sensible rules and guidance. But that shouldn’t involve shouting at them and humiliating them.

more from that review

Anyone interested in the education debate in the UK, a review by Sir Alan Wood of the safeguarding at the much lauded Mossbourne Academy Victoria Park should be a must read. Damning chscp.org.uk/wp-content/u...

Hi @tandfresearch.bsky.social who can I contact about submission portal errors preventing manuscript submission, as well as a Captcha error preventing support query submissions?

From top AI coding user globally to completely stop using it because it's so much harder to review AI code than write it myself. And they say all software developers jobs will be replaced by LLM. Life comes fast at you. 🤣

“There is a particular lack of understanding of neurodevelopmental needs and the impact of trauma and poor mental health on children’s learning and/or wellbeing. This results in children’s needs escalating and leading to reduced attendance or to them leaving school”

““They’re there because they want to be critical thinkers and politically engaged citizens.” […] they’re not paying for automation. They’re paying for mentorship, for dialogue, for intellectual relationships that can’t be outsourced to a chatbot”

“Calculators expanded reasoning; the printing press spread knowledge. ChatGPT, by contrast, doesn’t extend cognition—it automates it, turning thinking itself into a service. Rather than democratizing learning, it privatizes the act of thinking under corporate control. “

'If you want to be a whistleblower you have to be prepared to lose your job. I’m able to do what I’m doing here because I’m nobody' Old article but new to me. Loved it (Not just because it’s about a mid-life PGR from an IT background). Academic reality though: 😔 www.theguardian.com/science/2014...

We're organising a webinar on 18 February to mark the end of my @adr-uk.bsky.social Fellowship on school absences and crime. With talks from top-notch colleagues including @iainbrennan.bsky.social, @foxnic.bsky.social, Mark Mon-Williams, and many others! www.eventbrite.co.uk/e/from-schoo...

evergreen

Methodology

“institutions reward publication volume and citation metrics above everything else” “Papermills are merely the black-market version of the same sickness: a culture that rewards output, not validity; novelty, not truth; publication, not knowledge.”

“Overambitious […] underdelivered” Would be interesting to see an analysis of similarities with the Post Office IT scandal in terms of previous members of the board and management responses to staff/stakeholders raising concerns.

“ambitious plan to digitalize the courts […] an IT system, introduced as part of a £1bn project, has been plagued with technical faults - causing crucial information to go missing, be overwritten, or appear lost […] was known about for years before action was taken” www.bbc.co.uk/programmes/m...

How does a violence intervention program actually work? We interviewed 99 READI participants to find out. Key finding: They came for the job, stayed for the therapy. crimelab.uchicago.edu/resources/wh...