In the early stages of drafting a manuscript, it can be helpful to draw a little mascot (mousecot?) to cheer you on.

This gorgeous essay from the social-media-less Lily Hu is everything: on what it means to "do" philosophy amidst pressure to automate it (with lessons that travel much more broadly). Rhymes with brilliant insights from @biblioracle.bsky.social @janerosenzweig.bsky.social Ted Chiang & others

The UK AI pushback: 42% of people are deliberately limiting their AI use and 70% of people think it would be difficult or impossible to avoid being exposed to AI even if they wanted to. New report from us at @kings-dfi.bsky.social : www.kcl.ac.uk/news/over-40...

is this anything

I was interviewed for this Science Magazine piece on industry funding, philosophy of AI, and conflict of interest disclosures. www.science.org/content/arti...

Science coverage of this open letter with quotes from @bayesianboy.bsky.social @cailinmeister.bsky.social Craig Callender, Matthew Liao @naomioreskes.bsky.social and me. Journals, the ball is in your court now.

thank you Merino for saving us from r*naldo discourse. Now Belgium needs to step up and save us from the other discourse.

This, but also, importantly, the US government is divesting from higher ed and investing in the tech industry in a way that will increase reliance of research on industry funding. This will have long-lasting impacts on how objective our science is.

all of these statements can be true simultaneously

An important read.

Good thread (with some interesting counters in reply) on whether philosophy is making progress; Kevin argues that progress in philosophy typically happens when it spins out new fields. A good recent example is meta-science (spun out of phil of science). But what about 'AI ethics'?

Signed as soon as I read it. True, some philosophers will cheerlead for free (due to their tech optimism or with the expectation of funding later) and such behaviour is not covered by the proposed policy. But it would still be useful to be able to detect some cheerleading worth discounting.

A lot of philosophers are cheerleading for free though No conflict of interest, potentially just as problematic

As philosophy rushes to claim expertise on AI, AI companies rush to control the philosophical narrative. Unlike other disciplines, philosophy has no standards for reporting conflicts of interest. This petition endeavors to change that: docs.google.com/forms/d/e/1F...

home at last

If someone wants to rank preprints based on a probabilistic language generator, that's their business. I think it's more objectionable that author photos are posted alongside an apparently hallucinated quote attributed to them.

Two peer-reviewed paper credits per scientist per year remains my dream. No one needs more than two papers per author per year.

Important work. “The emerging literature blurs the line between integrity filtering, based on necessary but insufficient signals of validity like reproducibility of stated results or lack of fake citations, and epistemic filtering, which uses machine-detectable signals to judge scientific quality.”

My "I didn’t kill Imre Lakatos" T-shirt has people asking a lot of questions already answered by my shirt.

Welcome new followers. I am a scholar thinking about how AI/ML tools work in knowledge production. Please let me know if 1. you are in Philadelphia and you would like to be friends, 2. you run a philosophy or STS department and would like to make me permanently your colleague, or 3. you are a

An example of this with Perplexity. I asked it to remind me of an article abt relation b/w history of science and history of knowledge and it says yes it’s called ‘The History of Science and the History of Knowledge’ 2017 by Simon Schaffer. It is in fact by Lorraine Daston.

Hell, one of them decided that all citations are CS/math because DOIs have numbers in them. I have no idea why people would let these things near a bibliography. As is always the case, they do not perform literature review, they produce output in the form of what a lit review might look like

have reviewed a few AI citation correction packages, and they would all amplify the problem - if invented male authors are present, hit crossref, pick the first match. Authors are more exact matches than titles. "Better LLM tooling" does not fix this, you can't dig up.

It's worthwhile to document these impacts, because they are harmful. But also: "When utilized in literature review" ... for this AND MANY OTHER REASONS they shouldn't be used that way.

$38.5 billion dollars seems like an awful lot of dollars to lose.

By happenstance, I googled a philosophy colleague’s distinctive terminology today to find her papers on it. The first dozen search results were ALL her papers, or citations of them. The AI Overview didn’t mention her once. It falsely attributed her (non-tech) concept to some tech bros at Stanford.

Ther are more ramifications that specifically affect female authors. Clarivate is removing papers from Scopus if in a former name, claiming “essential maintenance on your profile, and this is the start of an intensive period of system upgrades and curatorial oversight. “

Want to counteract this problem? Here’s a reminder of this awesome tool from @dalezhou.bsky.social to promote citation diversity! github.com/dalejn/clean...

The irony of using an LLM-powered bot with a female-coded screen name to respond to an observation about gender bias in LLMs is not lost on me.

in 2019 I had a role training staff at a university and research library about ML + AI software applications and concepts almost all the existing literature on this subject stressed the problem of information bias, with a particular highlight on its minimisation of women & POC we KNEW. we WARNED

Ah so LLMs operate like senior male academics. Who could have envisioned that technology trained on existing texts could operate in a fundamentally conservative nature

I am begging people to read Weapons of Math Destruction by Cathy O'Neil or at least watch her Ted Talk.

has anybody written on LLMs and algorithm based decision making being pushed at a time non- straight white men were gaining more traction into decision making, allowing preservation of the older belief set?

Circa 2020 there was a flurry of studies documenting biased citation practices, including this one: www.nature.com/articles/s41...

Isn't @bayesianboy.bsky.social great? Super effective in moving the needle for all sorts of meaningful things, including getting the word out about this. No one I know is asking for preferential treatment; just fairness.

This happened to my colleague. Her work was hallucinated by the AI as belonging to a male philosopher who worked in ancient but hadn't written anything on the topic! And it removed a negation to reverse what her view was! Predictable - if the genAI prophets bothered to read what they plagiarize.

One way I caught students cheating via AI was that multiple assignments included references that attributed an article I wrote (& had assigned as reading) to Ed Yong. If only I wrote that well. It was weird though.

In spot tests I've done, I see this happen *all of the time*, such as with my own papers. The LLM would often rather hallucinate a male scientist than get my name right. 🤷‍♀️ It's kind of unsurprising that models trained unscrupulously on the entire internet are misogynistic

Something similar to. 3 just happened to me. One of my articles was credited to Yuval Noah Harari in a published paper! The authors say they are fixing this citation and other hallucinated ones...

I’m going to be honest, I write (for a living!) about how these tools work, including how they function (or fail to) in science and metascience applications; including, even, how biases infiltrate these applications. I could not have guessed that it was this bad or this blatant.

Generative AI is a tool designed to break labor, but to break some laboring groups harder than others

When utilized in literature review, LLMs consistently 1. fail to mention female authors in female-led literatures, 2. insist that men are more influential or more heavily cited when this is contradicted by objective citation counts, and 3. attribute women’s work to hallucinated male scholars.

Something I have noticed about attempting lit review with LLMs: Frontier models still hallucinate heavily when asked to produce citations to works in women-led literatures. When constrained to works that can be linked to, they simply re-attribute papers written by women to imaginary men.

Men — and women — continue to be mismeasured. Should be required reading for science-ML neophytes.

📢 We're hiring! Assistant Professor (tenure-track) in Philosophy at Lingnan University. Areas: ethics (value theory, normative, metaethics, applied) or philosophy of AI. Strong publication record required. Review begins 20 July 2026 philjobs.org/job/show/31557

Which is to say, there has been no unprecedented leap in the advancement of Anthropic's models. It's all pre-IPO BS and everyone is falling for it uncritically.

One of the better bits I’ve seen here. “Who benefits from this play?” is the vital question. And we need more Kirby in the conversation. Less debate of the putative merits, more analysis of power and goals. anthonymoser.github.io/writing/rhet...

Another week of historic developments in the smoke and mirrors sector.

The unprecedented leap in advancement of Anthropic’s frontier models is one of the number one global news item this week. No technical demonstration was necessary to achieve this effect. That there is successful marketing.

The export control directive banning state of the art language models from Anthropic is only counterproductive to the US AI industry if you don’t view it as yet another act in the current production of AI safety theatrics.