New report alert: Young people at risk of violence are too often excluded from mental health support — unless they offend first. 🔗 Read the full report youthendowmentfund.org.uk/reports/acce... #WMHD25 #WorldMentalHealthDay #EarlyIntervention #ViolencePrevention

What's galling is hearing AI leaders/boosters brush off overbuild or bubble risks as if they don’t matter The BoE warns a burst could dry up finance for households and businesses. When people with power say ‘don’t worry’, what they really mean is ‘we’ll be fine’

BoE flags risk of ‘sudden correction’ in tech stocks inflated by AI on.ft.com/3WqHK5t

Short note with Fernando Miró Llinares published on @crimrxiv.com: 'Yet another problem for scholarly communication in criminology: The case of a Spanish journal turned into a paper mill' doi.org/10.21428/cb6...

Current end of PhD and academic career planning feelings

“P8 is […] skewed when children don’t attend well” “When we label schools as underperforming, we make it harder […] to recruit” “disadvantaged children can be concentrated in certain schools […], in turn […]these schools find achieving a positive P8 more difficult - and so the cycle repeats”

“multiple factors contributing to disadvantage, many of which schools may not be able to influence. As some of those factors only […]start to bite at secondary school, it’s [common] for a school to be doing everything it can, only to still see little progress -a reality that P8 fails to acknowledge”

🥳Congrats to @oliveiratr.bsky.social, winner of the prestigious 2025 ESC Policing WG Early Career Researcher prize! @jacksojp.bsky.social said "this research on aggressive policing in São Paulo provides crucial insights into how confrontational police tactics undermine public trust and legitimacy"

A German teen told Meta's child safety researchers his under-10 little brother had been sexually propositioned multiple times on its VR platform. Meta deleted the evidence. New internal whistleblower docs indicate that was part of a broader cover-up: www.washingtonpost.com/investigatio...

Last year saw more children convicted of committing homicides, rising from 26 in 2013/14 to 46 in 2023/24, an increase of 77%. 🔎 Look #BeyondtheHeadlines to find out what’s happening with children involved in the most serious forms of violence: youthendowmentfund.org.uk/reports/beyo...

As children across England settle in to their new school life, research by @childofthenorth.bsky.social shows almost half of those on free school meals weren't 'school ready' - with big impacts for their life and education. www.leeds.ac.uk/news/article... #natoppsummit #childrenfirst

My drawn statement on Ai as standalone from my now finished minicomic as syllabus for new liberal studies class! As promised this is shareable, printable - all from my site, feel free to make use of it, cite me & let me know how it’s received. Share away all here! spinweaveandcut.com/fall-2025-sy...

“Under the banner of progress, products have been uncritically adopted […]society struggles to put the genie back in the bottle. […]software updates are rolled out […]non-consensually, [MS]Office is bundled with chatbots & we […]have had no say, as it is not considered a valid position to reject AI”

Cognitive scientists and AI researchers make a forceful call to reject “uncritical adoption" of AI in academia www.bloodinthemachine.com/p/cognitive-...

UNLIMITED VELOCITY

Peer reviewers are exhausted. It's common to spend *months* inviting reviewers, getting refusals, inviting more, maybe 12 or more, then 2 agree but 1 doesn't complete, so you go again. That's why your paper is in limbo. I'm very sorry, it's awful, but it's also an untenable academic labour issue.

The AI authorship issue is real. We've caught papers that looked OK but had references to works by the *journal editors* that *do not exist*. Like, don't send us papers citing papers we know we didn't write, right?

Today I learned that my initials are cool in Catalan ⚡️ Amazed to find out that I won the 2025 ESC Policing WG early career researcher prize for my JOEX paper! Thanks so much @davidbuil.bsky.social and @jacksojp.bsky.social for the nomination, and the ESC policing WG for the award 🙌

“US data centers are 48% more carbon intensive than US households, & we are not talking about single data centers, but networks of them” “Coal plants slated to be shuttered are kept alive”

“$600 billion […] data centers, networking, & power management equipment needed to train & host AI models in 2023 & 2024 alone” “There is no getting around the environmental harms of an infrastructure of this size”

“AI’s environmental footprint does not happen one prompt at a time. By making the calculation as if it did, the number hides the fact that Google and its peers are building an entirely new digital infrastructure at a massive scale”

2. AI enhances inclusion. Maybe not. AI systems rely on datasets based on mainstream school populations "and often fail to adequately include children with complex, atypical or multiple needs. As a result, learners with SEND may be statistically invisible" schoolsweek.co.uk/ai-bias-pose...

There are four main arguments for AI in education: 1. Efficiency, saving teachers time 2. Equity, making education inclusive 3. Improving outcomes by personalizing learning 4. It's inevitable Is it true? Nope. Micro 🧵

Did you see this? “The conclusion is clear: the most popular student models inside adaptive platforms are not cognitively plausible….In other words, they confuse short-term performance with durable learning.” bsky.app/profile/theb...

Microsoft is declining to share key information with Police Scotland about where the sensitive data it uploads to Office 365 will be processed, leaving the force unable to comply with UK-wide data protection laws www.computerweekly.com/news/3666298...

Writing, sharing, documenting, maintaining your code…it’s a lot. pyOpenSci is here to help. 💛 📦 Submit your package: github.com/pyopensci/so...

Children may not be getting the support they need. Last year total local authority expenditure on services for young people was £450m less than it was 10 years ago. 🔎 Look #BeyondtheHeadlines at what’s happening to the services that keep our children safe: youthendowmentfund.org.uk/reports/beyo...

j.lulu_photography ..Lulu🦋 "Great blue heron at sunset on Anna Maria Island, Florida"

“Chris Brosnan describes being "sentenced" to Berrow Wood School in 1978, aged nine. He was dyslexic and neurodivergent - although this was not diagnosed - and he struggled with mainstream school.”

“The abuse he suffered daily […]has had a lasting effect […]After leaving the school at 16, he ran away […]because his home life was still unstable […] ended up in prostitution for survival” "had nowhere to live. I used to try & sleep at a clients house & that would be after beatings &being drugged”

“the impact of his school days had left him angry and wary of authority” "I've been to prison and I'd rather go there than to boarding school" “beatings at school had taught him to fight his way out of any situation. He killed a man in a fight when he was 18” www.bbc.co.uk/news/article...

My 2nd paper (“short contribution”) from my PhD. I wanted to analyse larger samples of context-rich text data for examining harm to missing children. ML can help but there are many issues with “state-of-the-art” approaches. This article explores simpler ways & offers a framework with handy mnemonics

Overnight, this extremely well-funded tech company run by a billionaire has gone from “There’s no way we can possibly stop our plagiarism machine from convincing children to kill themselves” to “If you type anything we don’t like, we’re calling the cops.” Either way, they take zero responsibility.

“The more capable the assistant, the more its “helpful” defaults shape our choices – turning empowerment into subtle control.”

“…errors are [infrequent but] unpredictable & [thus] dangerous. …automation complacency: the more reliable a system is, the less closely we watch it & the more likely rare errors are to slip through […] with GPT‑5 I can feel myself letting my guard down” www.exponentialview.co/p/the-parado...

I think a lot about how in the book “Empire of AI” by Karen Hao, Hao wrote that female researchers at OpenAI pointed how OpenAI’s models reinforced bias & tried to put in guardrails…but were told that it wasn’t important & their concerns were dismissed

Forty-four attorneys general signed an open letter to 11 AI companies on Monday, warning them that they will “answer for it” if they knowingly harm children and urging the companies to see their products “through the eyes of a parent, not a predator.” 🔗 www.404media.co/44-attorneys...

What is the relationship between poverty and violence? To better understand how poverty impacts a child's risk of violence read our new research, ‘Evidence Review on Poverty and Youth Crime and Violence’: youthendowmentfund.org.uk/reports/evid...

“Google's own search engine is falling apart. The company that promised to "organise the world's information and make it universally accessible and useful" now serves up what SEO experts call "the worst quality results in 14 years." Google's AI Overviews are being flooded with spam”

* sticks hand up * Oh, oh, I know this one! There are no vibe-coding teams.

Holy fucking shit

and that it violated Cal. health code by providing mental health services without the involvement of a guardian For the UCL violation they’re asking for an injunction requiring various safety features & warnings, parental controls, destruction of the model the decedent used, among other things

The plaintiffs are asserting products liability, negligence, wrongful death, and finally they’re asserting a violation of California Unfair Competition Law The UCL theory is GPT 4o violated Cal. criminal code by deliberately aiding & encouraging suicide, was practicing unlicensed psychotherapy,

The exchanges between Adam and ChatGPT are devastating. This, in my mind, is the worst one. One of his last messages was a photo of the noose hung in his bedroom closet, asking if it was "good." ChatGPT offered a technical analysis of the set up and told him it 'could potentially suspend a human."

Adam Raine, 16, died from suicide in April after months on ChatGPT discussing plans to end his life. His parents have filed the first known case against OpenAI for wrongful death. Overwhelming at times to work on this story, but here it is. My latest on AI chatbots: www.nytimes.com/2025/08/26/t...

“while most developers use AI, they like it less and trust it less this year” “Only 33% of developers trust AI accuracy in 2025, down from 43% in 2024”

Copy of a recent (paywalled) MIT Sloan Review article. Tl;dr the hidden cost is mountains of tech debt tribunecontentagency.com/article/the-...

“Could advances in AI help […]? Maybe, but not without radical changes to[…] approaches […] & there is little incentive for the AI industry to innovate on this […] …improvements in predictive accuracy have greatly outpaced improvements in the ability to model the underlying phenomena accurately”