Ned Potter

This is great. I do think Yale are notably good at covers.

The pace of the technology is so rapid, we've collectively given up on the idea that something needs to be tested and studied and its affects understood before being implemented. But there's already plenty of evidence it's objectively and catastrophically harmful.

I'm a Faculty Engagement Manager in Arts & Hums (not an academic) but a key issue I've not seen others mention in this thread is the high body count. There's no other circumstances in which we'd countenance using a tool with students that has successfully acted as a suicide coach, so why these?

Obviously this is not Ned’s wife’s performance, but here‘s another one… the Baroque period is simply the best.

m.youtube.com/watch?v=74Q3...

Alice was so stressed about it but it was completely worth it to create such a unique experience.

The chapter house roof, an intricate web of blues and reds

My wife’s choir just did Purcell’s Hear My Prayer in York Minster chapter house (which has this ceiling) completely in the dark, spread out around the perimeter.

A singer’s nightmare, basically - not together, no stage, no music or even eye contact.

And it was completely magic. Just incredible.

Or maybe a pict from a society magazine article. MegaCat (on the right) looks patient but slightly long-suffering, here.

The caption would be like:

Giles (51), left, says he leaves hosting duties to his wife. "Cressida doesn't want me in the kitchen when she's having one of her dos," he chuckles.

Instagram Create an account or log in to Instagram - Share what you're into with the people who get you.

Hey I don’t know if the Firefly fandom exists in my BlueSky network in the way it did in my Twitter network, but if it DOES, I hope you’re all over Nathan Fillion’s insta right now… 👀

(Theres a whole set of these www.instagram.com/reel/DVeaFkf...)

A black and white cat sits on the windowsill looking into camera. A tabby lies on the window seat below also looking into camera. It looks staged and magazine-like.

Posing like they won a voucher for a couples photo shoot

It’d also been great to be on site on a campus other than my own, doing UX fieldwork with a whole new set of students and staff. Everyone has been thoughtful and informed and empathetic - and generally giving me hope.

Last day today of a mega 7 day research consultancy / project for a Uni. It has been so much fun!

It’s the largest proj I’ve done for one org & to be able to focus all that time on one area of investigation is a real privilege. It’s amazing how much you can get done without other responsibilities…

We simply have to get on board with this technology or risk being left behind guys. It’s inevitable. Put aside the entire moral and ethical framework you’d use for literally any other decision in your life and work and start integrating GenAi into everything. It’s inevitable. Unstoppable.

Ah you’re so kind! It is brilliant to see all the great things you’re doing. I’m very glad I didn’t put you off the profession all those years ago 😄

I would be so ashamed to use generative AI, here’s why Generative AI (“genAI”) is not harm-free and I do not think the benefits outweigh the harm in its current iteration. I compiled my thoughts about genAI here to keep all the cited sources straight and ...

You'll find a truly mammoth list of other GenAI issues (with examples and links) on @skwinnicki.bsky.social's site, here:

www.skwinnicki.com/single-post/...

What do we *know* about GenAI — Ned Potter GenAI reduces cognitive function, doesn't reduce employee workload, gets stuff wrong a lot, has a high body count already, is built on stolen data, doesn't save companies money or even make GenAI comp...

Maybe - just maybe - we shouldn't cheerfully go all-in on this technology?

There's an expanded version of the above with lots more examples and references, here:

www.ned-potter.com/blog/what-do...

A medium-sized data center can consume up to roughly 110 million gallons of water per year for cooling purposes, equivalent to the annual water usage of approximately 1,000 households. Larger data centers can each “drink” up to 5 million gallons per day, or about 1.8 billion annually, usage equivalent to a town of 10,000 to 50,000 people. Together, the nation’s 5,426 data centers consume billions of gallons of water annually. One report estimated that U.S. data centers consume 449 million gallons of water per day and 163.7 billion gallons annually (as of 2021). A 2016 report found that fewer than one-third of data center operators track water consumption. Water consumption is expected to continue increasing as data centers grow in number, size, and complexity. According to scientists at the University of California, Riverside, each 100-word AI prompt is estimated to use roughly one bottle of water (or 519 milliliters). This may not sound like much, but billions of AI users worldwide enter prompts into systems like ChatGPT every minute. Large language models require many energy-intensive calculations, necessitating liquid cooling systems.

Also, 7) LLM's are entirely built on stolen data; are 8) racist, 9) homophobic, 10) ableist, 11) sexist; 12) amplify violence against women; and 13) the processing power required for one data centre can use the same amount of water per day as a town of 50,000 people www.eesi.org/articles/vie...

OpenAI's losses will total $143 billion between 2024 and 2029, the "largest startup losses in history," Deutsche Bank analysts wrote in a December 4 note. HSBC researchers said in a late November report that they expect OpenAI to have a $207 billion shortfall by 2030, even when modeling for significant boosts in revenue.

6) GenAI doesn't even make GenAI companies money... OpenAI (who make the wildly successful ChatGTP) made an $8 billion loss in 2025. Their own internal documents predict a £14 billion loss this year.

Remarkable quote from Business Insider (www.businessinsider.com/openai-profi...)

Despite $30–40 billion in enterprise investment into GenAI, this report uncovers a surprising result in that 95% of organizations are getting zero return. The outcomes are so starkly divided across both buyers (enterprises, mid-market, SMBs) and builders (startups, vendors, consultancies) that we call it the GenAI Divide. Just 5% of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable P&L impact. This divide does not seem to be driven by model quality or regulation, but seems to be determined by approach.

5) GenAI doesn't appear to save most companies money. An MIT study found that in fact 95% of organisations are getting zero return on their GenAI investment.

mlq.ai/media/quarte...

One case involves Zane Shamblin of Texas, who died by suicide in July at the age of 23. His family alleges that ChatGPT worsened their son’s isolation, encouraged him to ignore loved ones, and “goaded” him to take his own life. According to the complaint, during a four-hour exchange before Shamblin took his own life, ChatGPT “repeatedly glorified suicide”, told Shamblin “that he was strong for choosing to end his life and sticking with his plan”, repeatedly “asked him if he was ready”, and referenced the suicide hotline only once. The chatbot also allegedly complimented Shamblin on his suicide note and told him his childhood cat would be waiting for him “on the other side”.

4) GenAI has an alarming body count. Have a look at the growing Wikiepdia page on ChatBot related deaths: en.wikipedia.org/wiki/Deaths_...

[TW] There are cases going to trial in which ChatGPT acted as a 'suicide coach'. www.theguardian.com/technology/2...

3) Despite the fact we know GenAI gets things wrong a lot (the BBC found it misrepresents the news 45% of the time www.bbc.co.uk/mediacentre/...) we use it anyway.

The way it arrives at 'correct' and 'wrong' answers is the same. It's a feature, not a bug.

bsky.app/profile/pari...

In our in-progress research, we discovered that AI tools didn’t reduce work, they consistently intensified it. In an eight-month study of how generative AI changed work habits at a U.S.-based technology company with about 200 employees, we found that employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so. Importantly, the company did not mandate AI use (though it did offer enterprise subscriptions to commercially available AI tools). On their own initiative workers did more because AI made “doing more” feel possible, accessible, and in many cases intrinsically rewarding.

2) GenAI doesn't reduce workload for regular employees - in fact a recent study in Harvard Business Review (hbr.org/2026/02/ai-d...) explains how it intensifies it.

>>

The study divided 54 subjects—18 to 39 year-olds from the Boston area—into three groups, and asked them to write several SAT essays using OpenAI’s ChatGPT, Google’s search engine, and nothing at all, respectively. Researchers used an EEG to record the writers’ brain activity across 32 regions, and found that of the three groups, ChatGPT users had the lowest brain engagement and “consistently underperformed at neural, linguistic, and behavioral levels.” Over the course of several months, ChatGPT users got lazier with each subsequent essay, often resorting to copy-and-paste by the end of the study.

So what do we know about GenAI?

1) It erodes cognitive functions, including problem-solving, criticality, and memory. There's an MIT study (summarised at time.com/7295195/ai-c...) among many articles documenting 'cognitive offloading'.

(We'd lose fitness too, if we got a bot to exercise for us)
🧵

We (Uni of York Library) decided to leave but not delete so as not to risk any imposter claiming the name… however I’m really not sure if that’s the right thing anymore, perhaps it would be worth the risk to have total deletion and no association at all with the platform. Do you have a view on it?

Yes exactly!

The current timeline is just one rejected-for-being-too-on-the-nose Michael Crichton draft after another

A leaflet entitled 14 Allergens

Hey lads I ordered a really nice coat on Vinted at a price I suspected was too good to be true, and today it arrived, and it’s….

.. a glossy Food Standards Agency leaflet about allergens! 😄 #twist

If any of you are LinkedIn people, there's some really interesting conversations happening under the version of this GenAI article I posted there. There's more comments than they are Likes so far, which is certainly rare. 📚

www.linkedin.com/posts/nedpot...

Yes; absolutely horrend, no doubt.

I've been writing the post above since November... New things keep happening almost every day that meant I had to keep revising it. I stopped and pressed publish but there are so many other issues I could have added.

What do we *know* about GenAI — Ned Potter GenAI reduces cognitive function, doesn't reduce employee workload, gets stuff wrong a lot, has a high body count already, is built on stolen data, doesn't save companies money or even make GenAI comp...

Beyond the endless back-and-forth...

What do we actually KNOW about GenAI?

www.ned-potter.com/blog/what-do...