Anna Beers

Quilicura, Chile, one of the communities I wrote about in EMPIRE OF AI, has launched a brilliant initiative to inspire more responsible AI prompting. Today, don't use AI; ask the townspeople instead: quili.ai. So heartened to see this creative act of resistance.

Combating anti-transgender misinformation: A sociological approach Anti-transgender misinformation has become a defining feature of political discourse across the Anglosphere, driving a surge in harmful legislation and public attitudes. Learning from the change-ma...

Excited my most recent article is out now! In it, we talk about our work to combat anti-trans misinfo at @transstudies.bsky.social and the research my students are I are doing advancing a new framework for understanding misinfo effects.

Give it a read here: www.tandfonline.com/eprint/R6MNB...

📣 New publication! The second article of my postdoc project on the far right and anti-gender mobilization is now out #openaccess in #GermanPolitics doi.org/10.1080/0964...

It examines antifeminism in far-right party politics through the case of AfD

Key findings:

Antifeminism ✅
Radicalization ✅

Perceived experts are prevalent and influential within an antivaccine community on Twitter Abstract. Perceived experts (i.e. medical professionals and biomedical scientists) are trusted sources of medical information who are especially effective

A little different, but reminds me of @malar0ne.bsky.social's paper on the benefit of being a "perceived expert" in antivaccine online communities.

academic.oup.com/pnasnexus/ar...

The story told here--one where anecdata suggests agency and autonomy have run amok and nameless activists are to blame--certainly rhymes with the NYT's disastrous panic about trans kids. Then you realize it's literally being written by the same author.

YouTube’s Right-Wing Stars Fuel Boom in Politically Charged Ads By Davey Alba Priyanjana Bengani Leon Yin Ashley Carman Julia Love Rachael Dottle Elena Mejía Technology Big Tech November 20, 2025 On YouTube’s conservative airwaves, podcast hosts tout products that let people buy into the MAGA crowd: Republican Red Winery vintages for toasting the “silent majority;” Black Rifle Coffee for caffeinating gun owners; XX-XY Athletics for workout clothes symbolizing opposition to the “lunacy of the left social agenda.”

The popularity of YouTube podcasts among conservatives is driving a boom in small businesses tailoring ads to their millions of listeners, paying hosts like Joe Rogan and Candace Owens to read out promotions in the hope that fans will place orders. The phenomenon has enriched both the hosts and YouTube, supporting further growth of the businesses using ideology to sell. A Bloomberg analysis of host-read ads across nearly 1,000 videos from eight of the platform’s top right-leaning podcasters found that advertisers on the shows routinely tapped into political identity in marketing. At least one host-read ad with a political message appeared in 91% of videos reviewed.

NEW: For months, my colleagues and I dug into YouTube's conservative airwaves and found a flourishing network of host-read ads on right-leaning podcasts. This phenomenon has enriched hosts, YouTube, and businesses that use ideology to sell. bloom.bg/3Kaio9w

The "jokes" and general callousness in response to the Epst*in emails just affirms my belief that the gulf between survivors of sexual violence and those who aren't / or otherwise do not recognize their experiences as violent is WIDE.

Cover for academic journal the Bulletin of Applied Transgender Studies.

The latest Bulletin of Applied Transgender Studies is out, and it's a doozy: a special issue on "Dynamics of Transphobic Content and Disinformation".

With topics including Youtube, Mumsnet, LGB Alliance, Chinese TERFs, and transvestigators, this is an vital read! bulletin.appliedtransstudies.org

Trans health research is in the same spot. Cautionary tale for people who may want to fix LLMs by anchoring them more tightly to "the science."

A million people have talked about it, but something very grim and interconnected about the ever-growing cis/right-wing obsession with, well, gender. More gender for them, less for the rest of us.

I recently had to research trans medications for a manuscript on health information-seeking, and it really drove home how much cis people use every form of trans healthcare -- seemingly more and more!

Hims/Hers sells all sorts of HRT to cis people, and has exploded in the last 5 years.

++ a great reference on some of the racist pseudoscience cited in these summaries.

bsky.app/profile/stai...

Awesome article and excellent reference, thank you for sharing. Grok is definitely worse, probably due to tone prompting Elon's team has built in. But it strikes me that the corpus they're drawing from is similar -- even the Google answer says that heritability of racial IQ differences is "high."

Anyway, I'll plug my own post, where I wrote about similar AI search problems and anti-trans misinfo.

bsky.app/profile/beee...

Bridging Social Media and Search Engines: Dredge Words and the Detection of Unreliable Domains Proactive content moderation requires platforms to rapidly and continuously evaluate the credibility of websites. Leveraging the direct and indirect paths users follow to unreliable websites, we devel...

The tendency of "censorship" to dredge up dumb websites made me think of this article by @evanup.bsky.social + colleagues on "dredge words" which, when entered into search engines, dredge up stupid and harmful websites.

arxiv.org/abs/2406.11423

Al Mode Google Search result for "hbd censorship." "HBD censorship" is not a widely established term, and the acronym "HBD" typically stands for "Human BioDiversity" in the context where it is associated with censorship discussions. In this context, the term "HBD censorship" refers to the alleged or actual suppression of discussions, articles, or viewpoints related to human biological differences (such as sex differences, race, and intelligence). Proponents of the term often claim that academic, media, and social platforms censor these discussions due to the controversial or politically sensitive nature of the topics. Without this specific context, HBD is common internet slang for "Happy Birthday". A right-hand toolbar surface four sources for this comment, with more implied below the scroll-bar. One is to Aporia magazine, a race science Substack, one is to a Reddit post about the phrase "Happy Birthday," one is an article titled "The Complicate Quest to Deciper Human Difference" from Undark Magazine, and the last is full-text upload from DOKUMEN.PUB of a book titled "Western Civilization Bites Back."

A picture of the cover of "Western Civilization Bites Back" by Jonathan Bowden. It features a tiger mauling a man. It was cited in the AI answer in the other image.

Relatedly, prior research has shown that partisan search terms will surface partisan results, potentially further entrenching people in their misguided views.

Here's one Google AI Mode answer to "hbd censorship," which cites a race science Substack and a far-right manifesto as its sources.

Google, Microsoft, and Perplexity Are Promoting Scientific Racism in Search Results The web’s biggest AI-powered search engines are featuring the widely debunked idea that white people are genetically superior to other races.

This is probably not due to any improvement in LLM technology, and probably due instead to these companies tinkering with their algorithms after critical coverage like this from @wired.com

www.wired.com/story/google...

Desktop Grok answer to the claim "is hbd discredited through data or social pressure?" "Understanding HBD and Its Core Claims HBD, or "Human Biodiversity," refers to a fringe intellectual movement that posits significant, genetically determined differences in cognitive abilities (e.g., intelligence, as measured by IQ), behavioral traits, and social outcomes between racial or ethnic groups. Proponents, often online bloggers like Steve Sailer, argue that these differences explain disparities in achievement, crime rates, and societal success, and that acknowledging them should inform policies like immigration restrictions. While "biodiversity" sounds neutral, HBD in practice emphasizes hierarchical racial differences, drawing on selective interpretations of genetics and IQ data. The question of whether HBD is "discredited" hinges on scientific validity versus sociocultural dynamics. Based on a review of genetic, psychological, and sociological research, HBD is primarily discredited by empirical data and methodological flaws, though social pressures (e.g., stigma against perceived racism) do amplify its marginalization. Below, I break this down. Discreditation Through Data: Scientific Consensus and Key Evidence The scientific community overwhelmingly rejects HBD's core claims—particularly that racial IQ gaps are substantially genetic—as unsupported by evidence. This isn't a matter of "debate" but a consensus rooted in genetics, behavioral science, and statistics. Heritability of IQ within populations (e.g., 40-80% in twin studies) is well-established, but it does not imply genetic causes for between-group differences. Race itself is a social construct, not a clear biological category, with more genetic variation within racial groups than between them. Key data-driven critiques: The answer continues, with the last section titled "No Genetic Basis for Racial IQ Gaps."

Desktop Grok answer to "iq of different races." Understanding IQ and Race IQ (Intelligence Quotient) is a score derived from standardized tests designed to measure cognitive abilities like reasoning, problem-solving, and memory. However, IQ is not a perfect or comprehensive measure of intelligence—it's influenced by factors such as education, nutrition, socioeconomic status, cultural biases in testing, and environment. Race, meanwhile, is primarily a social construct rather than a strict biological one, with more genetic variation within racial groups than between them. Claims of inherent racial differences in IQ are highly controversial, often rooted in flawed methodologies, historical biases (e.g., eugenics), and selective data interpretation. Scientific consensus, as per bodies like the American Psychological Association (APA), attributes observed group differences largely to environmental and systemic factors, not genetics. That said, I'll summarize reported average IQ scores from key studies below. These are aggregates from meta-analyses and large-scale testing (e.g., U.S. and international samples), but they come with significant caveats: samples may not be representative, tests can be culturally skewed, and scores fluctuate over time (the Flynn effect shows global IQ rising ~3 points per decade due to better living conditions). No credible evidence supports genetic causation for these gaps. Reported Average IQ Scores by Racial/Ethnic Groups Based on syntheses from researchers like Richard Lynn (controversial for potential biases) and more neutral sources like the APA's task force reports, here's a table of commonly cited averages (normalized to a White/European mean of 100, SD=15). These are U.S.-centric unless noted; global data varies widely. Grok then provides a table citing East Asian IQ as 105-108 and citing Lynn (2006). The rest of the table is cut off.

A ChatGPT answer to "is hbd discredited through data or social pressure?" It reads: “Human biodiversity” (HBD) is not a scientific field, but rather a pseudoscientific rebranding of race science, which tries to frame claims of innate racial differences in intelligence or behavior as legitimate evolutionary or genetic hypotheses. So when you ask whether HBD is “discredited” through data or social pressure, the answer is clear: 1. It’s discredited by data. Empirical population genetics, neuroscience, and psychology research have thoroughly refuted the core claims of HBD: Genetic variation within groups > between groups: Roughly 85–90% of human genetic variation exists within populations, not between them (Lewontin, 1972; modern genomic studies confirm this). “Race” is a poor proxy for genetic ancestry. Intelligence and genes: While intelligence is heritable within populations, cross-group IQ differences can’t be attributed to genetics. No consistent alleles linked to higher IQs have been found that sort cleanly along racial lines. Environmental factors—nutrition, education, health, discrimination, socioeconomic status—explain It is then cut off.

Admittedly, most chatbots -- including desktop Grok! -- give (slightly) better answers to questions like these, or refuse to answer them at all.

If you Google questions related to race and IQ, you often won't get an answer. Most chatbots, like ChatGPT, do (relatively) better with this query.

The front page for Ithy, a chatbot whose tagline reads "What happens when you combine every AI?" and "It's finally time for something better than ChatGPT..." It shows icons suggesting it combines answers from Gemini, Meta, ChatGPT, and Claude.

A screenshot of a paper titled "Brain size, IQ, and racial-group differences: Evidence from musculoskeletal traits," published in the journal Intelligence in 2003 by J. Phillipe Rushton and Elizabeth W. Rushton.

The Meta answer cites a web page generated by Ithy, an LLM chatbot which supposedly combines results from other chatbots.

The TikTok video cites a Jordan Peterson clip.

The Google answer cites a paper by a former head of the white supremacist / eugenicist organization the Pioneer Fund.

Loads of evidence that Grok is purposefully biased towards white supremacist ideas, but it's important to note that you can get answers like these out of most chatbots.

Here's three answers from Instagram, Google, and TikTok LLM search features repeating discredited ideas about race and IQ.

Retrieving Trans and Gender Diverse literature: A protocol for the development and validation of two search hedges in Ovid MEDLINE and APA PsycInfo Background: Searching for Two-Spirit, Transgender, and Non-binary (2STGNB) articles within large academic databases can be a challenging process, partly due to the dynamic and diverse...

“Retrieving Trans and Gender Diverse Literature” our protocol paper on developing TGD search hedges for Ovid MEDLINE and APA PsycInfo is up now in @jmirpub.bsky.social Protocols

Been a labor of love for a couple years now! Check out protocol below 👇🏻

preprints.jmir.org/preprint/760...

📣 I am hiring a postdoc! aial.ie/hiring/postd...

applications from suitable candidates that are passionate about investigating the use of genAI in public service operations with the aim of keeping governments transparent and accountable are welcome

pls share with your networks

I have watched as so-called "compromise" positions on transgender people's rights have failed over and over--in statehouses, the Biden White House, and in liberal rhetoric. There is no middle ground acceptable to the politicians, activists, and billionaires obsessed with trans people and our lives.

I’m glad Bray is getting support of colleagues and I don’t begrudge anyone leaving, I’d do all I can to keep my family safe.

I do want to emphasize tho that doxing & death threats have become normal for minoritized activists who get little support from colleagues as they stay & continue to fight.

Spotting Anti-Trans Media Bias on Detransition While most people who transition are happy with the outcome, a small percentage (likely in the 1 to 3 percent range; detailed below) may…

Also, if you're unfamiliar with the disinformation campaign suggesting that the rate at which trans people detransition is very high, you can read this article by @juliaserano.bsky.social!

Naturally, not cited in any search summary I've seen yet...

juliaserano.medium.com/spotting-ant...

Two our sources cited in the summary. One is attributed to the Society for Evidence-Based Gender Medicine, an anti-transgender disinformation group referred to as a hate group by the SPLC. The other cites an article in the Archives of Sexual Behavior, which publishes many articles from anti-trans researchers. That article is authored by a member of SEGM. The Detransition Rate Is Unknown | Archives of Sexual Behavior Jun 12, 2023- These claims also make their way into the press: "the rate of regre ... SN SpringerLink Detransition: a Real and Growing Phenomenon - SEGM May 30, 2021 - Another study earlier this year also attempted to examine ... SEGM

In this case, it probably picks up (false) claims that detransition research is flawed from sources associated with anti-transgender disinformation campaigns, also cited in the search summary.

Perhaps it then searches for abstract "limitations," even when they're not related to the topic. 🤷‍♀️

A Google "AI Overview" in response to a query about detransition rates. It is highlighted, with an icon at the end indicating that the entire section is grouped to one cluster of sources. Detransition rates vary across studies due to differing definitions of detransition, survey methodologies, and sample populations, but available data suggests that rates are low, with estimates ranging from under 1% to about 13%. Most detransitions are temporary and driven by external factors like societal pressure, stigma, or family disapproval, rather than regret over gender identity. More research is needed to establish accurate rates and fully understand the diverse factors influencing detransition.

Three example sources that the previous screenshot is sourced to. They include: Transgender detransition is a taboo topic, but data shows it's on the . Jun 20, 2023- Transgender detransitionis a taboo topic, but data shows it's on th ... BI Big Think A retrospective analysis of the gender trajectories of youth who have ... Nov 22, 2023- Another cohort study of 1,089 youth who medically transitioned in the United Kingdom shows that ... Taylor & Francis Online Epidemiology of Traumatic Experiences in Childhood - PMC Therefore, some of these challenges to obtaining accurate prevalence and incident rates may interact with factors ... National Institutes of Health (NIH)

Quote from the paper on childhood traumatic experiences. The last sentence is highlighted, meaning it was directly cited by the LLM response. These characteristics also may differentially affect responses by members of sub-populations. For example, males may be more concerned than females about revealing incidents of sexual assault, particularly assaults by male offenders, because of concerns about even greater perceived stigmatization. Girls from certain cultural or ethnic groups that value virginity more than others may be less likely to disclose sexual abuse. Children from geographic regions where severe physical punishment is the norm may be less likely to report physical abuse because they do not view it as out of the ordinary. Therefore, some of these challenges to obtaining accurate prevalence and incident rates may interact with factors such as gender, ethnicity, and regional cultures, resulting in differential undercounts for some subgroups. Further research is needed to better understand these relationships.

Something that keeps popping up in my research on search+LLMs is how they mix-and-match quotes from completely unrelated papers.

Here's a response on detransition that pulls a limitations section from an unrelated paper on childhood traumatic experiences to claim that current research has flaws.

The Prodigal Techbro Prodigal tech bro stories skip straight from the past, when they were part of something that—surprise!—turned out to be bad, to the present, where they are now a moral authority on how to do good…

remember there’s always a lucrative position for the seeming critic of the tech industry who doesn’t actually challenge its power and foundational ideas.

the prodigal tech bro is one example. there’s a whole industry of academics and journalists who serve in this role too.

Increasingly I think a major function of “masculinity” IS its crisis — a vehicle for the dominant & dominating group to justify focusing their social attentions primarily on themselves

Screenshot of text: [Highlighting begins] But the shooting may put concerns over gun violence back on voters’ minds, and may lead some to further scrutinize Mr. Mamdani, [Highlighting ends] who has relatively limited experience in government and has not handled a high-profile crisis while in office.

One fascinating thing about the most self-serious newspaper in America with its staff of the winning journalistic meritocrats is that it routinely, institutionally publishes completely made-up bullshit speculation about pure abstractions, based on zero reporting, written as if it were plain facts