Anna Beers
@beeeeeers.bsky.social
Postdoc @unituebingen.bsky.social, on the job market, studying social media, influencers, far-right politics, conspiracy science, disinformation, and monetization. six e's.
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.
Here's one Google AI Mode answer to "hbd censorship," which cites a race science Substack and a far-right manifesto as its sources.
November 11, 2025 at 2:49 PM
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.
Here's one Google AI Mode answer to "hbd censorship," which cites a race science Substack and a far-right manifesto as its sources.
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.
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.
November 11, 2025 at 2:49 PM
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.
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 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.
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.
November 11, 2025 at 2:49 PM
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.
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.
Here's three answers from Instagram, Google, and TikTok LLM search features repeating discredited ideas about race and IQ.
![A TikTok AI Search answer. In the middle are three videos. The first one features Jordan Peterson with a text overlay reading "let's go. Look at people who only have an IQ of 145." It has 27.7k likes. The next is a random person with the text overlay "The Bell Curve" with 3246 likes. The last is mostly cut off, and only has 19 lies. The summary reads:
Average IQ by Race
Summarized from sources [lists icon for Wikipedia and others].
Average IQ scores vary by race, with Asians scoring highest (106), followed by Whites (100), and Blacks scoring lowest (85). These differences are influenced by a complex mix of genetic, environmental,
and socioeconomic factors.
Key Findings
Global Averages:
East Asians: ~106
Whites: ~100
Blacks: ~85 (U.S.), ~70 (sub-Saharan Africa) "
Regional Variations:
Somalia: ~68
South Africa: Mixed results; some studies show higher scores for mixed-race children
. Historical Context:
IQ thresholds for disability classification have been racially biased (e.g., 70 for Whites, 80 for)
[The rest is cutoff.]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:tsuexertapdaedikorpy27ck/bafkreig5gzucsjejzrbwomfls5qrlnhpdh5vxzsehuasr7x44rwor4mqlu@jpeg)
November 11, 2025 at 2:49 PM
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.
Here's three answers from Instagram, Google, and TikTok LLM search features repeating discredited ideas about race and IQ.
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. 🤷♀️
Perhaps it then searches for abstract "limitations," even when they're not related to the topic. 🤷♀️
September 5, 2025 at 5:34 PM
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. 🤷♀️
Perhaps it then searches for abstract "limitations," even when they're not related to the topic. 🤷♀️
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.
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.
September 5, 2025 at 5:34 PM
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.
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.
anyway i can't (can) believe these apps got away with providing us an interface to rate the traits of other human beings on a mass scale
July 9, 2025 at 3:15 PM
anyway i can't (can) believe these apps got away with providing us an interface to rate the traits of other human beings on a mass scale
![For Black workers operating in the platform economy, racial profiling not only shapes which tasks they accept, but how they complete them. For example, DoorDash driver, Charles, told me: “I don’t wear hoodies, and I am always sure to have something that has the DoorDash logo on it, and I’ll wear my undergraduate university shirt, so they know I’m not a threat; I’m supposed to be here.” To prevent activating racialized emotions of fear and distrust, reinforced by stereotypes of Black youth as inherently criminal, Black platform workers operate under the assumption that they are always under surveillance.[20] If you are constantly under surveillance, what you wear and how you dress is incredibly important. According to the company’s website, catering bags with the official Uber logo range from $12 to $16, with entire safety kits that include vests ranging from $16 to $38.70.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:tsuexertapdaedikorpy27ck/bafkreianekwm5blkloueqa5423ovvcjzyjqfgmlnhmc7qb2gu4r62ftksy@jpeg)
Just finished this awesome essay from @citap.bsky.social postdoc Jaylexia Clark on @ssrc.org.
On "racialized emotions" in gig work like Uber and DoorDash, and how these apps surveil and punish Black gig workers' personal and emotional expression.
On "racialized emotions" in gig work like Uber and DoorDash, and how these apps surveil and punish Black gig workers' personal and emotional expression.
July 9, 2025 at 3:15 PM
Just finished this awesome essay from @citap.bsky.social postdoc Jaylexia Clark on @ssrc.org.
On "racialized emotions" in gig work like Uber and DoorDash, and how these apps surveil and punish Black gig workers' personal and emotional expression.
On "racialized emotions" in gig work like Uber and DoorDash, and how these apps surveil and punish Black gig workers' personal and emotional expression.
I'm at #ica25! Doing a high-density session on Saturday morning about how Google launders anti-trans orgs via their AI summaries.
HMU if you want to chat about trans/media things, or just going to war with the far-right in general.
HMU if you want to chat about trans/media things, or just going to war with the far-right in general.
June 13, 2025 at 3:32 PM
I'm at #ica25! Doing a high-density session on Saturday morning about how Google launders anti-trans orgs via their AI summaries.
HMU if you want to chat about trans/media things, or just going to war with the far-right in general.
HMU if you want to chat about trans/media things, or just going to war with the far-right in general.
If you go on Google Trends right now, and look at the "Rising" related queries for any low-volume search, you'll probably find kids looking up their homework questions.
..and why not, because Google will answer them lol.
..and why not, because Google will answer them lol.
May 22, 2025 at 3:43 PM
If you go on Google Trends right now, and look at the "Rising" related queries for any low-volume search, you'll probably find kids looking up their homework questions.
..and why not, because Google will answer them lol.
..and why not, because Google will answer them lol.
The second part focuses on how far-right Christian organizations smuggle their materials into search queries about transgender kids.
This one's more insidious. The AI summary massages a website promoting conversion therapy — making it sound like a reasonable, non-religious option for kids.
This one's more insidious. The AI summary massages a website promoting conversion therapy — making it sound like a reasonable, non-religious option for kids.
May 19, 2025 at 5:56 PM
The second part focuses on how far-right Christian organizations smuggle their materials into search queries about transgender kids.
This one's more insidious. The AI summary massages a website promoting conversion therapy — making it sound like a reasonable, non-religious option for kids.
This one's more insidious. The AI summary massages a website promoting conversion therapy — making it sound like a reasonable, non-religious option for kids.
As reported by the @splcenter.org, anti-trans pseudoscience groups often pay the open-access fees for researchers with friendly viewpoints.
It needs to be studied further, but this strategy seems to work for AI search. Many cited articles had their open access fees paid by an anti-trans group.
It needs to be studied further, but this strategy seems to work for AI search. Many cited articles had their open access fees paid by an anti-trans group.
![Screenshotted text.
One might imagine that algorithms like these could improve their answers to medical questions if they restricted themselves to citing the academic literature, rather than potentially partisan blogs—that is, sticking to the “science.” However, pseudoscientific disinformation campaigns are often laundered through the scientific literature itself, creating additional challenges for algorithmic systems hoping to do scientific boundary work. Prior reporting has documented how SEGM pays for open-access fees for authors publishing research congenial to their anti-transgender stances, so that their work will be freely accessible to the public instead of hidden behind paywalls. This also means, however, that the text of these articles is available to algorithmically generated search summaries like Google Search, which often cites their claims via publicly accessible text.[13] Two out of the three peer-reviewed scientific articles cited in the above summary on detransition had their open-access fees paid for by SEGM, of these, one suggests that existing research on detransition rates is flawed, and the other suggests it is higher than is widely reported. One of the articles was authored by someone whose only listed affiliation was SEGM itself.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:tsuexertapdaedikorpy27ck/bafkreibosvmqwvc4ddmc5aipn4awqu4hexq34cn4vtupz4eucbdgymojga@jpeg)
May 19, 2025 at 5:56 PM
As reported by the @splcenter.org, anti-trans pseudoscience groups often pay the open-access fees for researchers with friendly viewpoints.
It needs to be studied further, but this strategy seems to work for AI search. Many cited articles had their open access fees paid by an anti-trans group.
It needs to be studied further, but this strategy seems to work for AI search. Many cited articles had their open access fees paid by an anti-trans group.
It can be a little shocking which websites are integrated into AI search results.
This screenshot shows a query from the American College of Pediatricians, an ultra-right medical conspiracy group that opposes gay parenthood, being integrated into a question about gender-affirming care.
This screenshot shows a query from the American College of Pediatricians, an ultra-right medical conspiracy group that opposes gay parenthood, being integrated into a question about gender-affirming care.
May 19, 2025 at 5:56 PM
It can be a little shocking which websites are integrated into AI search results.
This screenshot shows a query from the American College of Pediatricians, an ultra-right medical conspiracy group that opposes gay parenthood, being integrated into a question about gender-affirming care.
This screenshot shows a query from the American College of Pediatricians, an ultra-right medical conspiracy group that opposes gay parenthood, being integrated into a question about gender-affirming care.
The first part is about how anti-trans pseudoscience orgs have camped out on certain search terms, like those relating to "detransition," to muddy the waters on gender-affirming care.
May 19, 2025 at 5:56 PM
The first part is about how anti-trans pseudoscience orgs have camped out on certain search terms, like those relating to "detransition," to muddy the waters on gender-affirming care.
Another weird AI search fuck-up today. Have been testing how Google's AI writes about detransition (spoiler: awfully). In this paragraph, in order to criticize a study it cites, it pulls in the limitations sections from.. completely unrelated studies.
Also, the paper it's criticizing isn't real 🤷♀️.
Also, the paper it's criticizing isn't real 🤷♀️.
March 4, 2025 at 12:25 AM
Another weird AI search fuck-up today. Have been testing how Google's AI writes about detransition (spoiler: awfully). In this paragraph, in order to criticize a study it cites, it pulls in the limitations sections from.. completely unrelated studies.
Also, the paper it's criticizing isn't real 🤷♀️.
Also, the paper it's criticizing isn't real 🤷♀️.
Original with alt text.
March 1, 2025 at 5:18 PM
Original with alt text.
Also revealing is Meta protecting race science on its platform. There's a lot of overlap between anti-trans influencers and "human biodiversity" eugenics influencers.
Again, either they're working backwards from the right-wing influencer oeuvre, or they're already immersed in it themselves.
Again, either they're working backwards from the right-wing influencer oeuvre, or they're already immersed in it themselves.
![Overall, the restrictions on claims of ethnic or religious supremacy has been eased significantly. The document explains that Meta now allows “statements of superiority as long as the statements do not refer to inferiority of another [protected characteristic] group (a) on the basis of inherent intellectual ability and (b) without support.” Allowable statements under this rule include “Latinos are the best!” and “Black people are superior to all others.” Also now acceptable are comparative claims such as “Black people are more violent than Whites,” “Mexicans are lazier than Asians,” and “Jews are flat out greedier than Christians.” Off-limits, only because it pertains to intellectual ability, is the example “White people are more intelligent than black people.”
[highlighted by me] But general statements about intellect appear to be permitted if they’re shared with purported evidence. For example, “I just read a statistical study about Jewish people being smarter than Christians. From what I can tell, it’s true!” It’s unclear if one would be required to link to such a study, or merely claim its existence.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:tsuexertapdaedikorpy27ck/bafkreihjyk6lefjjz7o63afn2utchvbkybbcjhx22q6p4slevm6oigzp6u@jpeg)
January 10, 2025 at 11:14 PM
Also revealing is Meta protecting race science on its platform. There's a lot of overlap between anti-trans influencers and "human biodiversity" eugenics influencers.
Again, either they're working backwards from the right-wing influencer oeuvre, or they're already immersed in it themselves.
Again, either they're working backwards from the right-wing influencer oeuvre, or they're already immersed in it themselves.
I've seen a lot of posts on Meta's new public content policy, but the examples of "allowed" statements in their internal guidelines are shocking.
Why write these down at all? My guess is they reverse engineer them from popular right-wing influencers' posts.
via www.platformer.news/meta-new-tra...
Why write these down at all? My guess is they reverse engineer them from popular right-wing influencers' posts.
via www.platformer.news/meta-new-tra...
January 10, 2025 at 6:29 PM
I've seen a lot of posts on Meta's new public content policy, but the examples of "allowed" statements in their internal guidelines are shocking.
Why write these down at all? My guess is they reverse engineer them from popular right-wing influencers' posts.
via www.platformer.news/meta-new-tra...
Why write these down at all? My guess is they reverse engineer them from popular right-wing influencers' posts.
via www.platformer.news/meta-new-tra...
Is there any world event that a New York Times writer would not chalk up to "polarized politics"
December 4, 2024 at 4:00 AM
Is there any world event that a New York Times writer would not chalk up to "polarized politics"
The white male bias of Twitter news influencers in my dissertation was so clear that I counted them out in a footnote.
Finding the same bias while researching news influencer monetization.
Finding the same bias while researching news influencer monetization.
November 19, 2024 at 12:03 AM
The white male bias of Twitter news influencers in my dissertation was so clear that I counted them out in a footnote.
Finding the same bias while researching news influencer monetization.
Finding the same bias while researching news influencer monetization.
Not great — Google AI Overviews integrating conservative outrage bait videos from YouTube to produce results for "teaching gender identity in the classroom." Results pretty innocuous here, but scary that these video transcripts are being gobbled up for search summaries.
November 15, 2024 at 3:31 AM
Not great — Google AI Overviews integrating conservative outrage bait videos from YouTube to produce results for "teaching gender identity in the classroom." Results pretty innocuous here, but scary that these video transcripts are being gobbled up for search summaries.
I've been looking at Google's AI overviews recently, and found a disturbing and utterly expected result on a query related to gender affirming care. The AI Overview approvingly cites the Society for Evidence Based Gender Medicine (SEGM), which the SPLC calls a "key hub of anti-LGBTQ+ pseudoscience."
![SPLC Quote: "Since its founding, members of SEGM have undertaken a global media and public policy blitz to challenge the affirming care model, advocate against gender-affirming care, and lend scientific credibility to legal claims against LGBTQ+ civil rights. Specifically, SEGM has helped foster resistance to the idea that adolescents can be capable of exerting agency over their own care. This strategy was first seen in the Bell v. Tavistock case, before carrying it over more in the U.S. SEGM, for example, has indicated its belief that exploratory psychotherapy should be a first-line treatment for gender dysphoric people age 25 and under.[11] The group’s endeavors are helped largely by its scientific façade, a general lack of information about its political activities and its members’ affiliations with the anti-LGBTQ+ far right, as well as its extensive connections and substantial personnel overlap with another powerhouse group founded in the U.K. called Genspect."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:tsuexertapdaedikorpy27ck/bafkreighrgdnl65rcu276bi3abeos2oll42ripfqmajhhm7s5ky3rlvngi@jpeg)
October 3, 2024 at 9:31 PM
I've been looking at Google's AI overviews recently, and found a disturbing and utterly expected result on a query related to gender affirming care. The AI Overview approvingly cites the Society for Evidence Based Gender Medicine (SEGM), which the SPLC calls a "key hub of anti-LGBTQ+ pseudoscience."