Will Berard 🫳🎤🫶

This is super interesting, but FFS, it needs 'linguistic and lexical representations of' inserted every time they mention a feeling.

Mathy maths can't feel pain, racist linear algebra can't be afraid, stochastic parrots do not get sad.

[2609.16247] The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It
https://arxiv.org/abs/2609.16247

#AI #LLMs #LLMPain

Student Loans in England (and Wales maybe? IDK) is fast becoming the "have you been mis-sold PPI?" of these years.

So, have you taken a student loan in the past years?

Write to your MP if you're on plan 2 (2021-2023): https://www.nus.org.uk/dont-freeze-our-futures-write-to-mp

Or on plan 5 (23 onwards):
https://www.nus.org.uk/dfof-write-to-mp-plan-5

#UKPol #UkEd #HigherEducation #StudentLoands #NUS #NationalUnionofStudents

"YOU WILL NEVER BELIEVE WHY HURTING MAKES YOU SAD!"

"Modern AI is fundamentally dependent on corporate resources and business practices, and our increasing reliance on such AI cedes inordinate power over our lives and institutions to a handful of tech firms.” (Whittaker, 2021)

Whittaker, M. (2021) ‘The Steep Cost of Capture’. Rochester, NY: Social Science Research Network. Available at: https://papers.ssrn.com/abstract=4135581 (Accessed: 6 July 2025).

#AI #BigTech

I'll just leave this here.

Generative AI Can Harm Teaching by Alp Sungu, Benjamin Lira, Angela Duckworth :: SSRN
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7007339

#AI #Teaching #Education @edutooters@a.gup.pe

RE: @openculture@toot.community

"recently, I have adopted the practice the Japanese refer at 'Tsundoku'"
-- Me, on a first date
#Books #bookstodon

You know that trend of running AI models through psychological tests run for humans?

To little surprise, this pre-print finds that instruments designed for human minds cannot be meaningfully applied to statistical engines for the production of text.

Meyer, J., Garcia, D. and Wulff, D.U. (2026) ‘Apparent Psychological Profiles of Large Language Models are Largely a Measurement Artifact’. arXiv. Available at: https://doi.org/10.48550/arXiv.2606.20205.

#AI #LLM #AIAlignment

Ok, but hear me out: what if we called the billion monkeys with typewriters 'agentic'?

#GenerativeAI #AI

“In a 1958 interview about the future of surveillance, Aldous Huxley commented that ‘these new techniques of propaganda will bypass the thought of rational side of man and appeal to the subconscious and deeper emotions and his physiology even and so making him actually love his slavery’ (Huxley, 11:04-11:18).” (Griffiths and Stockman, 2022, p. 78)

Griffiths, P. and Stockman, C. (2022) ‘Dark Patterns and Epistemic Ignorance: An Educational Crisis’

#Surveillance #Information #SocialMedia

Dig this: Google proposes to train model on synthetic data, for 'reasoning' in data-scarce domains.

Meaning AI decision making in, say healthcare, and the latter as, say particular combinations of genotype, demographics and certain disabilities and chronic diseases, maybe in comorbid combinations.

Designing synthetic datasets for the real world: Mechanism design and reasoning from first principles
https://research.google/blog/designing-synthetic-datasets-for-the-real-world-mechanism-design-and-reasoning-from-first-principles/

#AI #Google #SyntheticData

Don't get me wrong, I dig a MILF, but we need more love for the CCLILF (Childless Cat Lady ILF)

#MILF #CCLILF

“In reclaiming our science and our dignity, we need to reject the idea that fast is good or even neutral (Stengers 2018). When we show our students how slow science is not only possible, but necessary for a truly sustainable and liveable world, we empower critical scholars to resist fast science, and reject the concomitant AI infiltration.” ([Guest et al., 2025, p. 7]

Finally managed to draft a paper below the maximum recommended word count!

Title: Toward a Normative Theory of 'Generative AI' use in education

Body text: Don't.

“Large Language Models (LLMs) are so confounding for education in part because the very same use of technology can be hostile to familiar forms of education, yet educational in its own right.” (Bialystok, 2024, p. 105)

Bialystok, L. (2024) ‘Lines in the Sand: Originality and Cheating in the Age of ChatGPT’, Philosophy of Education Society [Preprint]. Available at: https://doi.org/10.47925/80.1.103.

#AI #Education #Philosophy #Bialystok

Why are we calling a machine 'intelligent', it's not like we call forklifts 'strong'?

Me:'See, I wanted a supervisor on the AI group of uni, cause I wanted to be inside the tent, pissing in'

Them: ...out? You mean out?

Me:...

This is what we are at risk of losing, folks.

#Memes #AI #University

Filed under 'the US will do X rather than control access to firerams'

"A Texas company is testing drones designed to stop school shooters using sirens, strobe lights and pepper spray. The idea isn’t new. In 2022, Taser-maker Axon proposed putting drones in schools — a plan that led most of its AI Ethics Board to resign."

Click Here: Taser drones in schools: Axon’s plan to stop school shootings

Episode webpage: https://play.prx.org/listen?ge=prx_8376_07c8b7d7-9703-4ac5-87bd-01424f96b4ac&uf=https%3A%2F%2Fpublicfeeds.net%2Ff%2F8376%2Fclickhere

Media file: https://tracking.swap.fm/track/sblTq32fyWAjsHzze2LG/mgln.ai/e/48/dts.podtrac.com/redirect.mp3/dovetail.prxu.org/8376/07c8b7d7-9703-4ac5-87bd-01424f96b4ac/v2-PODCAST_REPEAT_Axon_Drones_20260922.mp3

#Podcast #AI #Axon

Conferences fill me with energy, but it's pick-me energy

'After all, a witticism is merely the half-life of an emotion'
-- Henry Wotton, in _Dorian_ by Will Self

#quotes #bookstodon #WillSelf

"The strongest proponents of the SToE often ignore the stubborn fact that phones are global but their worst effects are strangely concentrated in the richest, English-speaking countries."

I think there are a couple of genocides in the global south that bet to differ, but sure.

Is the Smartphone Theory of Everything Wrong? A Comprehensive Investigation
https://www.derekthompson.org/p/is-the-smartphone-theory-of-everything

#Smartphones #DerekThompson #Technology #SocialMedia

The 'AI slowdown' bullshit is getting WILD!

Lawsuit says Anthropic, OpenAI, SpaceXAI and Google made illegal deal on AI slowdown - CBS News
https://www.cbsnews.com/news/ai-slowdown-lawsuit-openai-anthropic-google/

#Google #Anthropic #OpenAI #SpaceX #XAI #CBSNews

I have been doing audio engineering for the podcast of the University of Roehampton's sustainability team.

It features interviews with members of UR's community on student-led sustainability initiatives.

Available in all the places.
https://shows.acast.com/the-roey-sustainability-cafe

#Podcasts #Sustainability

Every.
Time.

Meta Tests Muse AI Agent Calls That Are Actually Made By Humans in a Call Center
https://www.404media.co/meta-tests-muse-ai-agent-calls-that-are-actually-made-by-humans-in-a-call-center/

#AI #Meta #404media

(getting vaccinated)

Nurse:... Sharp prick!

Me: rude but fair.