How latent and prompting biases in AI-generated historical narratives influence opinions

How latent and prompting biases in AI-generated historical narratives influence opinions

Abstract. Large language models (LLMs) can be used to persuade people on a range of issues, particularly through user-driven strategies such as personalizi

Relying on AI chatbots for historical facts can influence your political beliefs, new study shows

A recent study in PNAS Nexus suggests that reading history summaries generated by artificial intelligence can subtly shift your political opinions. Even when factually accurate, chatbots carry hidden biases that influence how people view the world.

The levers of political persuasion with conversational artificial intelligence

There are widespread fears that conversational artificial intelligence (AI) could soon exert unprecedented influence over human beliefs. In this work, in three large-scale experiments ( N = 76,977 participants), we deployed 19 large language models (LLMs)—including some post-trained explicitly for persuasion—to evaluate their persuasiveness on 707 political issues. We then checked the factual accuracy of 466,769 resulting LLM claims. We show that the persuasive power of current and near-future AI is likely to stem more from post-training and prompting methods—which boosted persuasiveness by as much as 51 and 27%, respectively—than from personalization or increasing model scale, which had smaller effects. We further show that these methods increased persuasion by exploiting LLMs’ ability to rapidly access and strategically deploy information and that, notably, where they increased AI persuasiveness, they also systematically decreased factual accuracy. , Editor’s summary Many fear that we are on the precipice of unprecedented manipulation by large language models (LLMs), but techniques driving their persuasiveness are poorly understood. In the initial “pretrained” phase, LLMs may exhibit flawed reasoning. Their power unlocks during vital “posttraining,” when developers refine pretrained LLMs to sharpen their reasoning and align with users’ needs. Posttraining also enables LLMs to maintain logical, sophisticated conversations. Hackenburg et al . examined which techniques made diverse, conversational LLMs most persuasive across 707 British political issues (see the Perspective by Argyle). LLMs were most persuasive after posttraining, especially when prompted to use facts and evidence (information) to argue. However, information-dense LLMs produced the most inaccurate claims, raising concerns about the spread of misinformation during rollouts of future models. —Ekeoma Uzogara , INTRODUCTION Rapid advances in artificial intelligence (AI) have sparked widespread concerns about its potential to influence human beliefs. One possibility is that conversational AI could be used to manipulate public opinion on political issues through interactive dialogue. Despite extensive speculation, however, fundamental questions about the actual mechanisms, or “levers,” responsible for driving advances in AI persuasiveness—e.g., computational power or sophisticated training techniques—remain largely unanswered. In this work, we systematically investigate these levers and chart the horizon of persuasiveness with conversational AI. RATIONALE We considered multiple factors that could enhance the persuasiveness of conversational AI: raw computational power (model scale), specialized post-training methods for persuasion, personalization to individual users, and instructed rhetorical strategies. Across three large-scale experiments with 76,977 total UK participants, we deployed 19 large language models (LLMs) to persuade on 707 political issues while varying these factors independently. We also analyzed more than 466,000 AI-generated claims, examining the relationship between persuasiveness and truthfulness. RESULTS We found that the most powerful levers of AI persuasion were methods for post-training and rhetorical strategy (prompting), which increased persuasiveness by as much as 51 and 27%, respectively. These gains were often larger than those obtained from substantially increasing model scale. Personalizing arguments on the basis of user data had a comparatively small effect on persuasion. We observe that a primary mechanism driving AI persuasiveness was information density: Models were most persuasive when they packed their arguments with a high volume of factual claims. Notably, however, we documented a concerning trade-off between persuasion and accuracy: The same levers that made AI more persuasive—including persuasion post-training and information-focused prompting—also systematically caused the AI to produce information that was less factually accurate. CONCLUSION Our findings suggest that the persuasive power of current and near-future AI is likely to stem less from model scale or personalization and more from post-training and prompting techniques that mobilize an LLM’s ability to rapidly generate information during conversation. Further, we reveal a troubling trade-off: When AI systems are optimized for persuasion, they may increasingly deploy misleading or false information. This research provides an empirical foundation for policy-makers and technologists to anticipate and address the challenges of AI-driven persuasion, and it highlights the need for safeguards that balance AI’s legitimate uses in political discourse with protections against manipulation and misinformation. Persuasiveness of conversational AI increases with model scale. The persuasive impact in percentage points on the y axis is plotted against effective pretraining compute [floating-point operations (FLOPs)] on the x axis. Point estimates are persuasive effects of different AI models. Colored lines show trends for models that we uniformly chat-tuned for open-ended conversation (purple) versus those that were post-trained using heterogeneous, opaque methods by AI developers (green). pp, percentage points; CI, confidence interval.

Redesigning algorithms to intervene on social norm misperceptions during a national election

For the first time in history, civic discourse commonly occurs in digital environments in which algorithms influence exposure to social information1,2. It is increasingly important to understand whether and how these algorithms affect political discourse3–5. Here we built custom feed-ranking algorithms with full control over their features, and randomly assigned 2,000 participants to use them for 8 weeks (before and after the 2024 US presidential election). We tested whether an engagement-based algorithm (used on major social media platforms6,7) amplifies intergroup, moralized and emotional (IME) information in ways that skew perceptions of social norms around political dialogue5,8, and whether it increased engagement with IME content and perceptions of partisan animosity (compared with a reverse-chronological feed9,10). We also developed and tested a ‘diversified extremity’ algorithm to reduce the influence of extreme users11–13 to improve the accuracy of social norm perception14–16 and reduce perceptions of partisan animosity. We found that engagement-based feeds amplified IME and toxic content relative to reverse-chronological feeds, with the largest increases in moral outrage and political content. Engagement-based feeds also reduced prescriptive norm perception accuracy (albeit in an unexpected direction) and increased perceived partisan animosity. However, they did not significantly alter users’ own engagement behaviours. The diversified extremity algorithm reduced IME and toxic content exposure, improved prescriptive norm accuracy, yet maintained comparable platform enjoyment—suggesting that reducing the influence of extreme users can curb algorithmic distortions without diminishing user experience.

The Dynamics of Distortion: How Successive Summarization Alters the Retelling of News

This work advances and tests a theory of how news information evolves as it is successively retold by consumers. Drawing on data from over 11,000 participants across ten experiments, the authors offer evidence that when news is repeatedly retold, it undergoes a stylistic transformation termed “disagreeable personalization,” wherein original facts are increasingly supplanted by opinions and interpretations with a slant toward negativity. The central thesis is that when retellers believe they are more (vs. less) knowledgeable than their recipient about the information they are relaying, they feel more compelled to provide guidance on its meaning and to do so in a persuasive manner. This enhanced motivation to guide persuasively, in turn, leads retellers to not only select the subset of facts they deem most essential but, critically, to provide their interpretations and opinions on those facts, with negativity being used as a means of grabbing their audience’s attention. Implications of this work for research on retelling and consumer information diffusion are explored.

The Dynamics of Distortion: How Successive Summarization Alters the Retelling of News

This work advances and tests a theory of how news information evolves as it is successively retold by consumers. Drawing on data from over 11,000 participants across ten experiments, the authors offer evidence that when news is repeatedly retold, it undergoes a stylistic transformation termed “disagreeable personalization,” wherein original facts are increasingly supplanted by opinions and interpretations with a slant toward negativity. The central thesis is that when retellers believe they are more (vs. less) knowledgeable than their recipient about the information they are relaying, they feel more compelled to provide guidance on its meaning and to do so in a persuasive manner. This enhanced motivation to guide persuasively, in turn, leads retellers to not only select the subset of facts they deem most essential but, critically, to provide their interpretations and opinions on those facts, with negativity being used as a means of grabbing their audience’s attention. Implications of this work for research on retelling and consumer information diffusion are explored.

LLM-generated messages can persuade humans on policy issues

The emergence of large language models (LLMs) has made it possible for generative artificial intelligence (AI) to tackle many higher-order cognitive tasks, with critical implications for industry, government, and labor markets. Here, we investigate whether existing, openly-available LLMs can be used to create messages capable of influencing humans’ political attitudes. Across three pre-registered experiments (total N = 4829), participants who read persuasive messages generated by LLMs showed significantly more attitude change across a range of policies - including polarized policies, like an assault weapons ban, a carbon tax, and a paid parental-leave program - relative to control condition participants who read a neutral message. Overall, LLM-generated messages were similarly effective in influencing policy attitudes as messages crafted by lay humans. Participants’ reported perceptions of the authors of the persuasive messages suggest these effects occurred through somewhat distinct causal pathways. While the persuasiveness of LLM-generated messages was associated with perceptions that the author used more facts, evidence, logical reasoning, and a dispassionate voice, the persuasiveness of human-generated messages was associated with perceptions of the author as unique and original. These results demonstrate that recent developments in AI make it possible to create politically persuasive messages quickly, cheaply, and at massive scale.

Biased AI writing assistants shift users’ attitudes on societal issues

Artificial intelligence (AI) writing assistants powered by large language models (LLMs) are increasingly used to make autocomplete suggestions to people as they write text. Can these AI writing assistants affect people’s attitudes in this process? In two large-scale preregistered experiments ( N  = 2582), we exposed participants writing about important societal issues to an AI writing assistant that provided biased autocomplete suggestions. When using the AI assistant, the attitudes participants expressed in a posttask survey converged toward the AI’s position. However, a majority of participants were unaware of the AI suggestions’ bias and their influence. Further, the influence of the AI writing assistant was stronger than the influence of similar suggestions presented as static text, showing that the influence is not fully explained by these suggestions, increasing accessibility of the biased information. Last, warning participants about assistants’ bias before or after exposure does not mitigate the attitude-shift effect. , Biased AI writing assistants shift people’s attitudes about societal issues; common interventions do not prevent this influence.

Biased AI writing assistants shift users’ attitudes on societal issues

Artificial intelligence (AI) writing assistants powered by large language models (LLMs) are increasingly used to make autocomplete suggestions to people as they write text. Can these AI writing assistants affect people’s attitudes in this process? In two large-scale preregistered experiments ( N  = 2582), we exposed participants writing about important societal issues to an AI writing assistant that provided biased autocomplete suggestions. When using the AI assistant, the attitudes participants expressed in a posttask survey converged toward the AI’s position. However, a majority of participants were unaware of the AI suggestions’ bias and their influence. Further, the influence of the AI writing assistant was stronger than the influence of similar suggestions presented as static text, showing that the influence is not fully explained by these suggestions, increasing accessibility of the biased information. Last, warning participants about assistants’ bias before or after exposure does not mitigate the attitude-shift effect. , Biased AI writing assistants shift people’s attitudes about societal issues; common interventions do not prevent this influence.

The political effects of X’s feed algorithm

Feed algorithms are widely suspected to influence political attitudes. However, previous evidence from switching off the algorithm on Meta platforms found no political effects1. Here we present results from a 2023 field experiment on Elon Musk’s platform X shedding light on this puzzle. We assigned active US-based users randomly to either an algorithmic or a chronological feed for 7 weeks, measuring political attitudes and online behaviour. Switching from a chronological to an algorithmic feed increased engagement and shifted political opinion towards more conservative positions, particularly regarding policy priorities, perceptions of criminal investigations into Donald Trump and views on the war in Ukraine. In contrast, switching from the algorithmic to the chronological feed had no comparable effects. Neither switching the algorithm on nor switching it off significantly affected affective polarization or self-reported partisanship. To investigate the mechanism, we analysed users’ feed content and behaviour. We found that the algorithm promotes conservative content and demotes posts by traditional media. Exposure to algorithmic content leads users to follow conservative political activist accounts, which they continue to follow even after switching off the algorithm, helping explain the asymmetry in effects. These results suggest that initial exposure to X’s algorithm has persistent effects on users’ current political attitudes and account-following behaviour, even in the absence of a detectable effect on partisanship.

The political effects of X’s feed algorithm

Feed algorithms are widely suspected to influence political attitudes. However, previous evidence from switching off the algorithm on Meta platforms found no political effects1. Here we present results from a 2023 field experiment on Elon Musk’s platform X shedding light on this puzzle. We assigned active US-based users randomly to either an algorithmic or a chronological feed for 7 weeks, measuring political attitudes and online behaviour. Switching from a chronological to an algorithmic feed increased engagement and shifted political opinion towards more conservative positions, particularly regarding policy priorities, perceptions of criminal investigations into Donald Trump and views on the war in Ukraine. In contrast, switching from the algorithmic to the chronological feed had no comparable effects. Neither switching the algorithm on nor switching it off significantly affected affective polarization or self-reported partisanship. To investigate the mechanism, we analysed users’ feed content and behaviour. We found that the algorithm promotes conservative content and demotes posts by traditional media. Exposure to algorithmic content leads users to follow conservative political activist accounts, which they continue to follow even after switching off the algorithm, helping explain the asymmetry in effects. These results suggest that initial exposure to X’s algorithm has persistent effects on users’ current political attitudes and account-following behaviour, even in the absence of a detectable effect on partisanship.

People are getting their news from AI – and it’s altering their views

Even when information is factually accurate, how it’s presented can introduce subtle biases. As large language models increasingly bring people the news, this bias is a looming problem.

People are getting their news from AI – and it’s altering their views

Even when information is factually accurate, how it’s presented can introduce subtle biases. As large language models increasingly bring people the news, this bias is a looming problem.

How LLMs Distort Our Written Language

Large language models (LLMs) are used by over a billion people globally, most often to assist with writing. In this work, we demonstrate that LLMs not only alter the voice and tone of human writing but also consistently alter the intended meaning. First, we conduct a human user study to understand how people actually interact with LLMs when using them for writing. Our findings reveal that extensive LLM use led to a nearly 70% increase in essays that remained neutral in answering the topic question. Significantly more heavy LLM users reported that the writing was less creative and not in their voice. Next, using a dataset of human-written essays that was collected in 2021 before the widespread release of LLMs, we study how asking an LLM to revise the essay based on the human-written feedback in the dataset induces large changes in the resulting content and meaning. We find that even when LLMs are prompted with expert feedback and asked to only make grammar edits, they still change the text in a way that significantly alters its semantic meaning. We then examine LLM-generated text in the wild, specifically focusing on the 21% of AI-generated scientific peer reviews at a recent top AI conference. We find that LLM-generated reviews place significantly less weight on clarity and significance of the research, and assign scores that, on average, are a full point higher. These findings highlight a misalignment between the perceived benefit of AI use and an implicit, consistent effect on the semantics of human writing, motivating future work on how widespread AI writing will affect our cultural and scientific institutions.

How LLMs Distort Our Written Language

Large language models (LLMs) are used by over a billion people globally, most often to assist with writing. In this work, we demonstrate that LLMs not only alter the voice and tone of human writing but also consistently alter the intended meaning. First, we conduct a human user study to understand how people actually interact with LLMs when using them for writing. Our findings reveal that extensive LLM use led to a nearly 70% increase in essays that remained neutral in answering the topic question. Significantly more heavy LLM users reported that the writing was less creative and not in their voice. Next, using a dataset of human-written essays that was collected in 2021 before the widespread release of LLMs, we study how asking an LLM to revise the essay based on the human-written feedback in the dataset induces large changes in the resulting content and meaning. We find that even when LLMs are prompted with expert feedback and asked to only make grammar edits, they still change the text in a way that significantly alters its semantic meaning. We then examine LLM-generated text in the wild, specifically focusing on the 21% of AI-generated scientific peer reviews at a recent top AI conference. We find that LLM-generated reviews place significantly less weight on clarity and significance of the research, and assign scores that, on average, are a full point higher. These findings highlight a misalignment between the perceived benefit of AI use and an implicit, consistent effect on the semantics of human writing, motivating future work on how widespread AI writing will affect our cultural and scientific institutions.

The Elusive Backfire Effect: Mass Attitudes’ Steadfast Factual Adherence

Can citizens heed factual information, even when such information challenges their partisan and ideological attachments? The “backfire effect,” described by Nyhan and Reifler (Polit Behav 32(2):303–330. https://doi.org/10.1007/s11109-010-9112-2, 2010), says no: rather than simply ignoring factual information, presenting respondents with facts can compound their ignorance. In their study, conservatives presented with factual information about the absence of Weapons of Mass Destruction in Iraq became more convinced that such weapons had been found. The present paper presents results from five experiments in which we enrolled more than 10,100 subjects and tested 52 issues of potential backfire. Across all experiments, we found no corrections capable of triggering backfire, despite testing precisely the kinds of polarized issues where backfire should be expected. Evidence of factual backfire is far more tenuous than prior research suggests. By and large, citizens heed factual information, even when such information challenges their ideological commitments.