LLMs believe false statements even after explicit warnings that they're false

Fine-tuning tests show "bias... toward confidently representing the claims as true."

There's Something Fundamentally Wrong With LLMs

LLMs aren't trained on the "vast majority of speech," experts warn, a major blind spot that could have sweeping consequences.

Leveraging Motivations to Curb Misinformation: Self-Affirmation Reduces the Appeal of Political Conspiracy Theories

Abstract: In an era where digital misinformation poses significant challenges to societal well-being, this study explores a novel approach to preserving information integrity by addressing the motivational underpinnings of conspiracy theory engagement. As conspiracy theories proliferate online, traditional fact-checking and debunking strategies often prove ineffective due to the self-reinforcing nature of conspiracy theories. This research investigates whether a priori self-affirmation interventions can reduce individuals’ propensity to engage with conspiracy theories by preemptively fulfilling the ego-protective function these theories often serve. Using a randomized experiment ( N = 451), this study finds that participants who completed a self-affirmation task were less likely to read politicized conspiracy theories, and those who chose to read a conspiracy theory reported greater feelings of affirmation than those who did not. Thus, bolstering individuals’ self-integrity may offer an opportunity to mitigate the appeal of politicized conspiracy theories.

Well-intentioned obscenity

An LLM lied about me at the office and I'm cranky about it. That interaction makes me think a lot about how LLMs are becoming normalized, though, and what we're looking at in terms of their role in human-to-human interactions in the future.

LLMs are bullshitters. But that doesn't mean they're not useful.

Note: This is a personal essay by Matt Ranger, Kagi's head of ML In 1986, Harry Frankfurt wrote On Bullshit. He differentiates a lying from bullshitting: Lying means you have a concept of what is true,...

Hey ChatGPT, write me a fictional paper: these LLMs are willing to commit academic fraud

Mainstream chatbots presented varying levels of resistance to deliberate requests for fabrication, study finds

Hey ChatGPT, write me a fictional paper: these LLMs are willing to commit academic fraud

Mainstream chatbots presented varying levels of resistance to deliberate requests for fabrication, study finds.

Capturing our Attention

What are the harms of misinformation? It’s widely believed that it causes false belief, including false beliefs in quite bizarre claims. I argue that this threat is greatly exaggerated: while plaus...

Short Dialogues with AI Reduce Belief in Antisemitic Conspiracy Theories

A groundbreaking study supported by ADL found that fact-based dialogues with a trained AI chatbot durably reduced antisemitic conspiracy beliefs and increased favorability toward Jews.

Dan Shipper 📧 on Twitter / X

this is true and is a big reason why you don’t need to be a highly technical researcher to use LLMs in surprising and novel ways https://t.co/TuxNzXzToU— Dan Shipper 📧 (@danshipper) July 27, 2025

CFP: Religion, Politics, and Cognitive Warfare: Information, Interpretation, Conspiracy, and the Struggle for Reality

The Journal for Cultural and Religious Theory (JCRT) invites proposals for an online conference entitled Religion, Politics, and Cognitive Warfare: Information, Interpretation, Conspiracy, and the Struggle for Reality. The proceedings will be considered for publication in a special issue of the JCRT. This conference investigates how beliefs, paranoia, and conspiratorial modes of knowing shape a contemporary cognitive battlespace in which actors struggle to define truth, authority, and reality itself. Drawing on Michel Foucault's Lectures on the Will to Know and Julia Kristeva's This Incredible Need to Believe (2024), we focus on practices of selective fact use and "truth-selecting" that construct alternative epistemic orders while claiming privileged access to what is "really" going on. Rather than treating conspiracy thinking as merely political or psychological, the conference foregrounds its religious dimensions: faith in hidden powers, moral dualism, apocalyptic expectation, and hermeneutical struggle over revelatory access to the real. Conspiracy cultures routinely reproduce theological structures of knowing—visions of salvation and corruption, truth and deception, initiation and blindness. At the same time, new forms of "cognitive warfare" reframe these religious energies within state and platform attempts to govern perception, attention, and trust. Artificial intelligence, algorithmic media, and strategic information campaigns do not only deliver messages; they create digital ecologies in which suspicion becomes faith-like and revelation is continual. The conference organizers are seeking proposals of high academic quality that take a reflective and analytical approach to both general and specific topics with international appeal or focus. We encourage contributions from scholars of religious studies, philosophy, communication, sociology, security studies, psychology, media, and related fields. Submissions should engage religion as a dynamic force—conceptually, historically, or materially—within the cognitive battlespaces of our time. We encourage contributions from a spectrum of perspectives, approaches, and methodologies. We are not interested in papers that simply rehearse or promote particular conspiracies, or that use the conference as a platform for calling out disliked groups. We are not interested in polemics, jeremiad, or de facto advocacy pieces. Instead, we welcome theoretically informed and methodologically rigorous work that takes the religious and epistemic dimensions of cognitive conflict seriously as objects of critical inquiry.

Science flags paper that found AI chatbots help debunk conspiracy theories

Science has issued an expression of concern for a highly publicized study looking into whether conversations with AI chatbots could convince conspiracy theorists to abandon their beliefs. The move …

"AI Psychosis" in Context: How Conversation History Shapes...

Extended interaction with large language models (LLMs) has been linked to the reinforcement of delusional beliefs, attracting clinical and public concern. Yet most empirical work evaluates model...

The Liar’s Dividend: Can Politicians Claim Misinformation to Evade Accountability?

This study addresses the phenomenon of misinformation about misinformation, or politicians "crying wolf"' over fake news. Strategic and false claims that stories are fake news or deepfakes may benefit politicians by helping them maintain support after a scandal. We posit that this benefit, known as the "liar's dividend," may be achieved through two politician strategies: by invoking informational uncertainty or by encouraging oppositional rallying of core supporters. We administer five survey experiments to over 15,000 American adults detailing hypothetical politician responses to stories describing real politician scandals. We find that claims of misinformation representing both strategies raise politician support across partisan subgroups. These strategies are effective against text-based reports of scandals, but are largely ineffective against video evidence and do not reduce general trust in media. Finally, these false claims produce greater dividends for politicians than alternative responses to scandal, such as remaining silent or apologizing.

Does Developing a Belief in One Conspiracy Theory Lead a Person to be More Likely to Believe in Others?

ABSTRACT The monological belief system model suggests that—for at least a subset of people—developing a belief in one conspiracy theory will cause them to be more likely to believe in others. This model has been influential in the literature, but its core causal hypothesis has never been credibly tested. We therefore tested it in two longitudinal studies. Study 1 used a sample from New Zealand and Australia ( N  = 498), with 7 monthly waves. Study 2 (preregistered) used a sample from New Zealand, Australia and the United Kingdom ( N  = 978), with 13 monthly waves. We applied random intercept cross‐lagged panel models, permitting a credible causal identification strategy, albeit we cannot rule out time‐varying confounds. We find that increased belief in a conspiracy theory at one wave did (on average) predict increased belief in other conspiracies at the next wave, although the estimated coefficients were small.