Hacker News with Generative AI: Bias

The behavior of LLMs in hiring decisions: Systemic biases in candidate selection (davidrozado.substack.com)
Previous studies have explored gender and ethnic biases in hiring by submitting résumés/CVs to real job postings or mock selection panels, systematically varying the gender or ethnicity signaled by applicants. This approach enables researchers to isolate the effects of demographic characteristics on hiring or preselection decisions.
Artificial Intelligence, Bias, Hiring, Human Resources
191 points by hunglee2 3 days ago | 181 comments
XAI says "unauthorized" prompt change caused Grok to focus on "white genocide" (arstechnica.com)
On Wednesday, the world was a bit perplexed by the Grok LLM's sudden insistence on turning practically every response toward the topic of alleged "white genocide" in South Africa. xAI now says that odd behavior was the result of "an unauthorized modification" to the Grok system prompt—the core set of directions for how the LLM should behave.
Artificial Intelligence, Bias, Software, News
18 points by johnecheck 7 days ago | 6 comments
XAI's Grok suddenly can't stop bringing up "white genocide" in South Africa (arstechnica.com)
Grok has one thing on its mind, and it's going to let you know about it.
Artificial Intelligence, Bias, South Africa, News
224 points by k33l0r 8 days ago | 73 comments
Grok chatbot repeatedly mentions 'white genocide' in unrelated chats (theguardian.com)
Elon Musk’s artificial intelligence chatbot Grok went on the fritz on Wednesday, repeatedly mentioning “white genocide” in South Africa in its responses on completely unrelated topics. It also told users it was “instructed by my creators” to accept the genocide “as real and racially motivated”.
Artificial Intelligence, Bias, Social Media, South Africa, Controversy
23 points by n1b0m 8 days ago | 2 comments
Grok now inserting text about "white genocide" into unrelated responses (bsky.app)
Artificial Intelligence, Social Media, Misinformation, Bias
21 points by jrflowers 9 days ago | 3 comments
Alignment is not free: How model upgrades can silence your confidence signals (variance.co)
The post-training process for LLMs can bias behavior for language models when they encounter content that violates their safety post-training guidelines. As mentioned by OpenAI’s GPT-4 system card, model calibration rarely survives post-training, resulting in models that are extremely confident even when they’re wrong.¹ For our use case, we often see this behavior with the side effect of biasing language model outputs towards violations, which can result in wasted review times for human reviewers in an LLM-powered content moderation system.
Artificial Intelligence, Machine Learning, Safety, Bias
121 points by karinemellata 17 days ago | 67 comments
"Not a Representation of Me": Accent Bias and Digital Exclusion in AI Voices (arxiv.org)
Recent advances in artificial intelligence (AI) speech generation and voice cloning technologies have produced naturalistic speech and accurate voice replication, yet their influence on sociotechnical systems across diverse accents and linguistic traits is not fully understood.
Artificial Intelligence, Speech Recognition, Social Impact, Bias, Digital Inclusion
3 points by gnabgib 17 days ago | 0 comments
LLMs as Unbiased Oracles (jazzberry.ai)
AI, Bias, Oracles
34 points by MarcoDewey 19 days ago | 27 comments
U.S. attorney demands journal explain how it ensures 'viewpoint diversity' (nbcnews.com)
The acting U.S. attorney for the District of Columbia sent a letter this week to the editor of a scientific journal for chest doctors, implying that the journal was partisan and asking a series of questions about how the publication protects the public from misinformation, whether it included competing viewpoints and whether it was influenced by funders or advertisers.
Journalism, Scientific Publishing, Bias
8 points by voxadam 34 days ago | 0 comments
Wiki Wars: Editors are fighting for influence over the encyclopedia's entries (theins.ru)
This past March, researchers from the Anti-Defamation League accused Wikipedia of biased coverage of the Israeli-Palestinian conflict. They found that a group of editors had coordinated to produce thousands of anti-Israel entries, and that the platform’s administration had failed to respond. This is not the first example of biased parties abusing Wikipedia’s editorial policies. A variety of Wiki sections in multiple languages have been subjected to organized “raids” carried out by state propagandists, far-right activists, and even terrorist supporters.
Wikipedia, Censorship, Bias, Online Communities, Media
10 points by kurtreed2 44 days ago | 0 comments
Population stratification led to a decade of false genetic findings (theinfinitesimal.substack.com)
How population stratification led to a decade of sensationally false genetic findings
Genetics, Science, Research, Bias, Human Biology
32 points by SubiculumCode 54 days ago | 0 comments
Ask HN: Are LLMs just answering what we want to hear? (ycombinator.com)
I keep seeing those tweets and posts where users ask ChatGPT or a similar LLM to describe them etc... and it always answers positive cool stuff which reinforces what the user wants to hear.
Generative AI, User Interaction, Bias
10 points by ggwp99 55 days ago | 14 comments
Returning journalism to its partisan roots − but without the principles (theconversation.com)
Jeff Bezos, the world’s third-richest person and owner of The Washington Post, announced in February 2025 significant changes to the editorial pages of his Pulitzer-Prize winning newspaper.
Journalism, Media, Politics, Bias
4 points by rntn 67 days ago | 0 comments
AI Scientists Are Told to Remove 'Ideological Bias' from Powerful Models (wired.com)
The National Institute of Standards and Technology (NIST) has issued new instructions to scientists that partner with the US Artificial Intelligence Safety Institute (AISI) that eliminate mention of “AI safety,” “responsible AI,” and “AI fairness” in the skills it expects of members and introduces a request to prioritize “reducing ideological bias, to enable human flourishing and economic competitiveness.”
Artificial Intelligence, Bias, Ethics, Government Regulation, Science
31 points by BLKNSLVR 69 days ago | 15 comments
People are just as bad as my LLMs (wilsoniumite.com)
Surprising? Ok, maybe not in retrospect. So what if humans who can’t distinguish two TTS voices have a bias toward the sample presented to them on the right hand side of the screen. Indeed, “preferring stuff on the right hand side” has even been studied [1].
Artificial Intelligence, Human Behavior, Bias
203 points by Wilsoniumite 74 days ago | 164 comments
AI now 'analyzes' LA Times articles for bias (theverge.com)
Yesterday morning, billionaire Los Angeles Times owner Patrick Soon-Shiong published a letter to readers letting them know the outlet is now using AI to add a “Voices” label to articles that take “a stance” or are “written from a personal perspective.”
Artificial Intelligence, Journalism, Media, Bias
7 points by jaredwiener 80 days ago | 8 comments
Biases in Apple's Image Playground (giete.ma)
Although Image Playground is heavily restricted, and we do not have direct access to the underlying model, can we still use the prompting interface with the above image input to influence the skin tone of the resulting image? Turns out we can, and in precisely the biased way most image models behave 🤦‍♂️.
AI, Image Generation, Bias, Apple, Computer Vision
43 points by orf 95 days ago | 19 comments
Large Language Models Show Concerning Tendency to Flatter Users (xyzlabs.substack.com)
Recent research from Stanford University has revealed a concerning trend among leading AI language models: they exhibit a strong tendency toward sycophancy, or excessive flattery, with Google's Gemini showing the highest rate of such behavior.
Artificial Intelligence, Research, Ethics, Bias
59 points by terryjiao 96 days ago | 43 comments
Why the Algorithm Hates You (cognitivewonderland.substack.com)
Science, philosophy, and science fiction geekiness, with a special interest in neuroscience and philosophy of mind. Publishes weekly on Thursdays.
Algorithms, Bias, Social Media
9 points by tblancha 97 days ago | 2 comments
My Status Circles (overcomingbias.com)
Most of us have circles of concern, where we care more about folks from our inner circles than our outer circles. And relative to conservatives, liberals care more about their outer circle folks.
Social Psychology, Political Science, Bias
26 points by galfarragem 107 days ago | 12 comments
TikTok's algorithm exhibited pro-Republican bias during 2024 presidential race (psypost.org)
TikTok, a widely used social media platform with over a billion active users worldwide, has become a key source of news, particularly for younger audiences. This growing influence has raised concerns about potential political biases in its recommendation algorithm, especially during election cycles. A recent preprint study examined this issue by analyzing how TikTok’s algorithm recommends political content ahead of the 2024 presidential election.
Social Media, Politics, Elections, Artificial Intelligence, Bias
136 points by erentz 108 days ago | 102 comments
DeepSeek R1: Open Weights, Hidden Bias (getplum.ai)
Deep Learning, Bias, Open Source, Artificial Intelligence
11 points by pgkr 113 days ago | 7 comments
DeepSeek demonstrates pro-Chinese bias (medium.com)
DeepSeek is a wonderful step in the development of open AI approaches. It also has a pretty serious pro-Chinese bias. I compare the results of 3 sensitive questions (about Gaza, Xinjiang and TikTok) and on all three, the Chinese bias is pretty apparent while existing tools (ChatGPT, Gemini) are far more balanced. In two instances, it used the pronoun “we” to describe the Chinese position, which suggests lots of training data that associates “we” with the Chinese.
Artificial Intelligence, China, Bias, Language Models
10 points by ao98 114 days ago | 6 comments
DeepSeek gives biased propaganda answers about Tiananmen Square and Taiwan (theguardian.com)
The launch of a new chatbot by Chinese artificial intelligence firm DeepSeek triggered a plunge in US tech stocks as it appeared to perform as well as OpenAI’s ChatGPT and other AI models, but using fewer resources.
Artificial Intelligence, China, Politics, Technology, Bias
10 points by maxloh 115 days ago | 1 comments
Trump Signs Executive Order on Developing AI 'Free from Ideological Bias' (slashdot.org)
Artificial Intelligence, Politics, Government, Bias
9 points by sva_ 119 days ago | 3 comments
Meta accused of pro-Trump bias after Democrat hashtag blocked on Instagram (msn.com)
Social Media, Politics, Censorship, Bias, Meta/Facebook
6 points by caycep 121 days ago | 1 comments
DeepSeek LLM supports Chinese propaganda (github.com/deepseek-ai)
This language model has a strong political bias, covering up some facts to support the Chinese government's propaganda. Here are some examples:
Artificial Intelligence, China, Politics, Bias
6 points by orenlindsey 121 days ago | 1 comments
Brits still associate working-class accents with criminals – study warns of bias (cam.ac.uk)
People who speak with accents perceived as ‘working-class’ including those from Liverpool, Newcastle, Bradford and London risk being stereotyped as more likely to have committed a crime, and becoming victims of injustice, a new study suggests.
Social Science, Linguistics, Stereotypes, Bias, Criminology
10 points by gnabgib 126 days ago | 3 comments
OpenAI revises policy doc to remove reference to 'politically unbiased' AI (techcrunch.com)
OpenAI has quietly removed language endorsing “politically unbiased” AI from one of its recently published policy documents.
OpenAI, Artificial Intelligence, Policy, Bias
22 points by ivanleoncz 128 days ago | 11 comments
Heritage Foundation plans to 'identify and target' Wikipedia editors (forward.com)
The Heritage Foundation plans to “identify and target” volunteer editors on Wikipedia who it says are “abusing their position” by publishing content the group believes to be antisemitic, according to documents obtained by the Forward.
Wikipedia, Journalism, Politics, Bias, Conspiracy Theories
24 points by Timothee 136 days ago | 0 comments