I resigned from ExxonMobil today. I spent the last three years doing fossil fuel extraction research at both BP and ExxonMobil. Neither company is acting responsibly. They are racing to maximize extraction while the window to act closes, and gambling with our lives. More below.

I used to love this time of year, hearing from students applying to work in my lab. No more. Every day a fresh batch of nearly identical emails with nonsensical remixes of the text on my lab website. It makes me so sad.

Organizing and facilitating a workshop / panel this Wednesday, with @lilianedwards.bsky.social, @adamlopez.bsky.social, @datacontroversies.bsky.social If you're around, come check it out. We'll be focused on how to bring legitimacy to the technology embedded in everything. Sign up below.

Bringing Legitimacy to "AI Sovereignty"

What is Digital or AI sovereignty? What makes the technologies integral to our daily lives legitimate? How do we achieve better futures?

eventbrite.com

Leaders who can't be bothered to do their jobs, & want to be able to blame someone else for the decisions they want to take, hire consultants who care nothing about education, research, or people. Universities end up wrecked, lives are ruined, only consultants benefit. Again & again & again.

Margot Finn@eicathomefinn.bsky.social · 3mo ago

'The Herald reported in April that PA Consulting had been contracted to manage an "academic size and shape" exercise at the College of Arts, Humanities and Social Sciences (CAHSS).' 1/2

Resurfacing online to mention: my book on the history of Speech Rec, NLP, and data-driven AI is coming out in a couple of weeks! Just got my author's copy from @uchicagopress.bsky.social and it finally feels real! press.uchicago.edu/ucp/books/bo...

Divination Engines

A revealing and surprising origin story, showing how attempts to render human speech and language computable led from the era of big data to today’s AI. Since the advent of computers, society has fant...

press.uchicago.edu

While you're cooking in the heat this week, do bear this in mind.

We all know that senior management of the UK's universities are struggling with an impossible financial situation imposed by the government. But the way they have been handling this situation has also shown that the sector is led by some really callous & horrible people.

Plashing Vole@plashingvole.bsky.social · 3mo ago

Wake up; open email; discover my humane, caring boss has been forced out with no notice; meeting called to announce the next stage of own redundancy process this afternoon. Feels like everything's collapsing round our ears.

NLP reviews suck! But why is not clear– @aclrollingreview.bsky.social provides *a lot* of guidance for how to review, but in that, first principles get lost. So @adamlopez.bsky.social and I have written down some of our thoughts on first principles.

The Missing First Principles of Reviewing for ACL

Zeerak Talat & Adam Lopez, University of Edinburgh

medium.com

A year ago: We're building data centers everywhere to run you out of your job and double your power bill. It's so you can look at fake videos of talking dogs on your phone Now: Why are people so hostile about the data centers

Which is the fundamental problem in AI. It doesn't actually serve the people using it. In fact, often AI serves people who want to replace or undermine the people using it. It really is that simple.

"ChatGPT, the most heavily used AI service, gave wrong information in 46% of its answers, including making up an expenses scandal, giving inaccurate replies on voter eligibility rules and getting the date of the election wrong by two months." www.theguardian.com/technology/2...

ChatGPT and other AI bots made huge errors before Scottish election, study finds

Exclusive: Electoral Commission calls for new controls as Demos finds tools made up fake scandals, invented candidates or gave wrong date

theguardian.com

Tech Policy Press fellow Petra Molnar highlights the AI Resist List: a global database documenting acts of resistance to the AI industry. From legal challenges and worker organizing to artistic interventions, the project seeks to challenge the “scale at all costs” development of AI.

The World Is Already Resisting AI. Now, There is a List to Prove It.

Petra Molnar spotlights the launch of the AI Resist List, documenting global challenges to AI expansion.

techpolicy.press

As one of the authors of this thought I’d share a bit about how we got here and how you can do what we’ve done at you institution. Over the two years I’ve been at the University of Edinburgh I’ve grown increasingly concerned by fellow academics uncritically using LLMs, especially OpenAIs’s ChatGPT.

Emily M. Bender@emilymbender.bsky.social · 6mo ago

So a great thing about contracts is that they usually have and end date and/or means of ending them. If your university or other organization bent the knee to OpenAI or any of the other chatbot providers, when a contract is up, there is an opportunity for pressure: forms.gle/ZGn5brEGU4t9...

The Epstein case I'm afraid reveals something many academics don't want to see. Universities can be rapacious money machines, obsessed with money and status for their own sakes, heedless of risk, oftentimes willing to turn a blind eye when they shouldn't. apnews.com/article/epst...

Professors in the Epstein files say they hoped friendship would lead to research funding

A new trove of documents released by the Justice Department reveals that Epstein’s reach into academia was wider than previously known.

apnews.com

I'm prepping my spring NLP course, but I've not been following NLP research too closely lately. Luckily this primer popped up on LinkedIn, so I'm asking the hive mind... POLL: How crucial is it for me to cover Apeezies & Cogematic? a) Very b) Less crucial than Shangriant d) It's too cutting edge

The budget levy on international fees seems to confirm that the UK government is, for reasons that defy economic logic, determined to break the backs of its own world-class universities just as the US government does the same. Now why would a government want to destroy the modern university I wonder

which I know from personal inspection. What it had was the biggest (n-gram) language model anyone had yet built. @nsaphra.bsky.social et al. have a nice paper on this analogy. arxiv.org/abs/2311.05020

First Tragedy, then Parse: History Repeats Itself in the New Era of Large Language Models

Many NLP researchers are experiencing an existential crisis triggered by the astonishing success of ChatGPT and other systems based on large language models (LLMs). After such a disruptive change to o...

arxiv.org

From the University of Woeville: Alternative FAQs for British academics facing redundancy juliecupples.wordpress.com/2025/11/02/f...

A sobering read for anyone working on machine translation.

Brian Merchant@bcmerchant.bsky.social · last yr.

Since 2023, translators and interpreters have seen work dry up, rates plummet and their jobs reduced to editing AI-generated output. Some are leaving the field, others are considering bankruptcy. All despite any major upgrade in translation quality. This is how AI is killing translation work:

@nsaphra.bsky.social is awesome and so much fun to work with. Go work with her!

Naomi Saphra@nsaphra.bsky.social · 2y ago

Life update: I'm starting as faculty at Boston University @bucds.bsky.social in 2026! BU has SCHEMES for LM interpretability & analysis, I couldn't be more pumped to join a burgeoning supergroup w/ @najoung.bsky.social @amuuueller.bsky.social. Looking for my first students, so apply and reach out!

To understand the future of AI, take a look at the failings of Google Translate theconversation.com/to-understan...

To understand the future of AI, take a look at the failings of Google Translate

Nearly 20 years after it was launched, machine translation is still a long way from replacing translators.

theconversation.com

Ruby Ostrow, Adam Lopez LLMs Reproduce Stereotypes of Sexual and Gender Minorities https://arxiv.org/abs/2501.05926

📕 ⬇️ My thesis on 🚫unargmaxable outputs is online! Check it out if you want to learn more about how output layers constrain what neural networks can and cannot predict 👉 era.ed.ac.uk/handle/1842/...