Ben Williamson

The UK’s AI training ambition deserves better than a course directory - LSE Impact A lack of connection to the UK government’s own assessment of AI skills requirements &a wider vision beyond work undermines the potential of the AI Skills Hub.

🗣️"What we found was a mixture of existing training content loosely connected by the topic of AI."

#AISkills #AITraining

SCALE at Stanford U (recently did a big trial of Tutor CoPilots based on GPT-4) and CompSci dept at Tartu U (Estonia has huge national deal for ChatGPT in schools).

"Some sources even went to completely unrelated links that weren’t written by the person whose work they were supposedly an example of, potentially indicating that the suggestions Grammarly’s AI offers with one person’s name may be based on a different person’s work."

California colleges spend millions on faulty AI chatbots Community colleges are spending millions on AI chatbots that students say often give inaccurate answers. Many might see upgrades soon.

“In testing by CalMatters, they [the chatbots] often answered general questions correctly but struggled with more specific ones. East Los Angeles College’s bot couldn’t even correctly name its own president.”

Setting aside many things here, this doc is also full of arguable claims presented without evidence. Perhaps, for example, the best way for educators to help students "harness AI" is to make sure we teach them first to think/write/analyze/have ideas themselves. Then they can harness away.

... and it possibly being seriously deleterious for students, educators and educational institutions in the long term ... 🤮

Text from OpenAI article on its educational programs: "Agency does not emerge from basic AI use alone. Students must progress from simple tasks to deeper applications such as studying, building, creating, coding, and managing agents. But among college-age users, we see a widening global “capability overhang,” defined as the gap between what AI tools can do and how people actually use them. Even advanced student users still operate roughly 90% to 99% below how power users of ChatGPT are engaging with our tools. Educational institutions play a central role in closing that gap. Faculty and educators can help students harness AI’s full potential by embedding authentic AI use cases into coursework—assignments that use AI and mirror real professional work, such as analyzing a market, designing a product concept, evaluating a policy trade-off, or building a simple agent"

Shouldn't let OpenAI puff about education get to me, but its claim students need AI skills for future jobs and "agency", then letting slip it really means making students into GPT "power users", and making that an educational imperative, *and* planning to measure it ... 😦 openai.com/index/ai-edu...

OpenAI, Google and Microsoft are all laying claims to expertise in educational research. They're appealing to their own AI, data, and scale capacities to do this, and wrapping it in the language of learning and cognitive sciences. It's a big tech takeover of educational research imo 🥶

OpenAI diagram about its educational research workflow.

Here is how OpenAI is reimagining educational research. It's basically GPT accesses all student records and "over-time signals" for constant tracking. It's total student surveillance for its model improvement.

Whether you think AI may have good applications for education or not, it's clearly *not right* for global AI vendors to be controlling the production of evidence about their own products.

LearnLM outperformed other AI models in a recent technical study. LearnLM is our set of AI models and capabilities that infuses learning science into Gemini and the products it powers, including Search, YouTube and Classroom.In a recen…

Google is working hard to be an educational research authority on AI in education too. It's "learning science" to put chatbots in classrooms according to their own studies blog.google/products-and...

New tools for understanding AI and learning outcomes OpenAI introduces the Learning Outcomes Measurement Suite to assess AI’s impact on student learning across diverse educational environments over time.

Under no circumstances should we allow OpenAI to become the self-authorized educational research evidenve source that it is trying to be. Vendors must not be research authorities. openai.com/index/unders...

"AI" is Yesterday's News I gave this talk yesterday to the Massachusetts Teachers Association, as part of the MTA Retired First Wednesday Speaker Series. Thank you very much for inviting me to speak to you today. I’m honored...

"'AI' marks one of the most pressing labor issues that everyone in every profession is facing, particularly since the technology – or certainly those funding and forcing its adoption – is resoundingly anti-worker," writes Audrey Watters.

I enjoyed her takedown of Sal Khan's read of the Diamond Age etc. Do we read these texts differently or do they not actually read them?

Grammarly Is Offering ‘Expert’ AI Reviews From Your Favorite Authors—Dead or Alive The tool, offered by the recently-rebranded company Superhuman, gives feedback based on the work of famous dead and living writers—without their permission.

Read the full piece (including much longer quotes from me) by the perpetually excellent @milesklee.bsky.social below!

www.wired.com/story/gramma...

Grammarly Offering Manuscript Reviews by AI Versions of Recently Deceased Professors The Grammarly "Expert Review" feature uses AI to provide feedback on papers using the name and work of real professors, dead or alive.

Daily reminder that calling ai dead labor and stolen labor is literal.

When You’re an ‘Expert’ Reviewing Students’ Work on Grammarly — But You Didn’t Know It The platform offers users AI-generated feedback under the names of scholars by drawing on their accessible publications. Some academics are unsettled.

I stand by what I said in this article: "Look, nobody enters academia to get rich, but I think business models that are built on basically extracting other people’s work without attribution or consent are fundamentally predatory".

cc @janerosenzweig.bsky.social @hpsvanessa.bsky.social @ceaubin.com

Just went in to Grammarly and pasted in this article that I wrote for the Globe last year about finding my mother's book in LibGen and thinking about it as training data and clicked on expert review. A thread: Grammarly chose the experts. writinghacks.substack.com/p/for-there-... /1

I have seen a lot of cursed stuff in my time in academia but this is among the *most* cursed.
Grammarly is generating miniature LLMs based on academic work so that users can have their writing ‘reviewed’ by experts like David Abulafia, who died less than two months ago.

Cheating machine or powerful assistant? The AI anxieties of a trainee teacher The long read: I was a newcomer, negotiating all of usual classroom difficulties for the first time. Throwing AI into the mix felt like downing a coffee in the middle of a panic attack

Cheating machine or powerful assistant? The AI anxieties of a trainee teacher

Private equity and venture capital investors alike see education in terms of assets for acquisition, control and - mostly - future returns. And that will change the nature of the things they have made into their return-yielding assets.

Text reads: Conclusion The education sector presents a compelling investment opportunity shaped by geographic concentration in an evolving market. Education's unique positioning as economic infrastructure underpins its investment appeal. The sector's role as a GDP multiplier, where each 1% increase in education expenditure corresponds to 21.3% GDP growth, distinguishes it from traditional service sectors through both direct returns and broader economic value creation. This dual nature provides defensive characteristics whilst creating substantial consolidation opportunities across a market projected to reach $8tn by 2030.

Private equity is now a key actor - discursively, organizationally, financially - in shaping the education sector. But its gaze sees only future returns from financial calculations on asset control. That's not how most education workers see the value of the sector
www.macfarlanes.com/insights/102...

Text reads: Public sector funding constraints present significant opportunities for private capital to drive technology-enabled solutions that address growing skills gaps6 whilst delivering more scalable, cost-effective, and interactive educational experiences. Unlike public institutions constrained by budget cycles and procurement processes, private capital can move quickly to implement technological solutions and respond to evolving market needs.7

Third, from a private equity perspective, public sector austerity is an opportunity for massively scalable private technosolutions to be rolled out at pace. It's a common narrative of the investor as a "saviour" of the public sector's failings...

Text reads: Education's role in human capital development creates sustainable competitive advantages that transcend economic cycles. Research shows that each extra year of schooling leads to an approximately 9% annual increase in a person’s earnings,and the associated human capital generates positive externalities and spillover effects that benefit entire economies through enhanced productivity and innovation.5 The sector’s inelastic demand also provides defensive investment qualities particularly valuable during periods of economic uncertainty.

Second, private equity views education as "human capital" investment. So education is configured as *valuable* from an investors' vantage point in terms of "positive externalities" and "spillover effects" into "productivity". This feels like an image of education that AI would amplify...

Text reads: Why education? Education’s investment appeal stems from its fundamental role as economic infrastructure. The global education market is projected to expand from $6tn in 2022 to $8tn by 20303, representing a compound annual growth rate (CAGR) of 3.7%, a trajectory that reflects sustained market expansion despite economic uncertainties.

How the private equity industry views education is really alien to how educators see it. Yet these investors probably have a bigger say in how the future of education will unfold.

First, education is a *market* worth $trillions *in the future* making it a good speculative bet for investment...

That Won't Work, Bro Announcing my new ed tech consultancy.

Maybe I shouldn't be, but I'm surprised people take this Alpha School thing at all seriously. Does no one remember Rocketship or AltSchool or the other attempts at this digitally automated schooling? Time to re-up this from the AltSchool days. www.insidehighered.com/blogs/just-v...

Einstein & The Rise of Nuisance Tech AI may be one of the most transformative innovations of our lifetime, but not when it is treated as spam.

Einstein isn’t really a threat to academic integrity so much as it is another distracting piece of marketing hype. It isn’t clear if it was a hoax, a failed app launch, or piece of tech posing as performance theater. AI agents represent challenges to online learning open.substack.com/pub/marcwatk...

Generic white dude who programs @westbynoreaster Why then did you take down the “Einstein” chatbot? 1:22 AM · Feb 26, 2026 · 67 Views Advait Paliwal @advaitpaliwal · 16h Cease and desist Generic white dude who programs @westbynoreaster · 15h Really? Presumably from Canvas/Instructure, right? Advait Paliwal @advaitpaliwal · 7h Due to the name Einstein

In utterly DELIGHTFUL news, Adwait Paliwal, the desi techbro behind the cheatbot Einstein AI which claimed it could log into Canvas and do/turn in your assignments for you has been forced to take down his website.
He'll likely be back, and there are others like him in abundance, sadly.

"Universities are not peripheral to Scotland's economy, rather they are anchor institutions that can generate £28.30 billion in annual economic impact, support 34,000 jobs through research funding, and create employment multiplier effects..."

digitalpublications.parliament.scot/ResearchBrie...

UK universities turn to Duolingo to test foreign students’ English Try questions from the test used by more than 50 UK higher education institutions

Universities reportedly 'lack the resources to undertake research' into which tests are most effective. Pull your socks up universities: research is fundamental to what you do, and this is not rocket science.