arXiv subjects

Martin Saveski

Publications and source records attributed to Martin Saveski.

18 recordsLinked to original sources

arXiv2026

How Human Feedback Shapes AI-generated Community Notes

Community Notes, a bridging-based crowd-sourced fact-checking system, has emerged as a new mechanism for moderating misleading information on social media and has been adopted by major platforms including X, Facebook, Instagram, Threads, and TikTok. Since its introduction, there has been an open question about what role AI could play in scaling and optimizing the system. Recently, X extended its Community Notes system by introducing Collaborative Notes: notes initially drafted by an LLM and iteratively refined based on feedback from human contributors. In this work, we systematically analyze the complete corpus of 19,146 collaborative notes and 211,850 instances of human feedback. First, we develop a taxonomy of human suggestions for improving AI-generated note drafts and find that suggestions involving factual corrections and additional context are most likely to be incorporated, while subjective policy judgments rarely are. Second, we examine changes in helpfulness across versions of collaborative notes and find that human feedback leads to more helpful notes, with the greatest impact coming from suggestions that challenge the main claim in the previous draft, particularly when submitted by more active contributors. Finally, we find that although collaborative notes improve through human feedback, they reach helpful status and are shown on the platform at lower rates than human-only or AI-only notes, with limited human participation emerging as a key bottleneck. Nevertheless, rather than serving as a weaker substitute, collaborative notes tend to play a complementary role, predominantly targeting posts that do not attract human-only or AI-only notes. Our analysis provides an initial description of efforts to use AI to improve crowdsourced content moderation in a real-world moderation system and outlines pathways for future improvements to such features.

cs.CY↗

arXiv2026

Detecting and Enhancing Intellectual Humility in Online Political Discourse

Intellectual humility (IH)-a recognition of one's own intellectual limitations-can reduce polarization and foster more understanding across lines of difference. Yet little work explores how IH can be systematically defined, measured, evaluated, and enhanced in spaces that often lack it the most: online political discussions. In this paper, we seek to bridge these gaps by exploring two questions: 1) how might preexisting levels of IH influence future expressions of IH during online political discourse? and 2) can online interventions enhance IH across different political topics and conversational environments? To pursue these questions, we define a codebook characterizing different dimensions of IH and intellectual arrogance (IA) and have researchers use it to annotate several hundred Reddit posts, which we then use to develop and validate a classifier to support IH analysis at scale. These tools subsequently enable two key contributions: i) an observational data analysis of how IH varies across different political discussions on Reddit, which reveals that more/less IH environments tend to contain future posts of a similar nature, and ii) a randomized control trial evaluating strategies for nudging discussion participants to demonstrate more IH in their posts, which reveals the possibility of enhancing IH in online discussions across a range of contentious topics. Our findings highlight the possibility of measuring and increasing IH online without necessarily reducing engagement.

cs.CY↗

arXiv2026

The Prosocial Ranking Challenge: Reducing Polarization on Social Media without Sacrificing Engagement

We report the first direct comparisons of multiple alternative social media algorithms on multiple platforms on outcomes of societal interest. We used a browser extension to modify which posts were shown to desktop social media users, randomly assigning 9,386 users to a control group or one of five alternative ranking algorithms which simultaneously altered content across three platforms for six months during the US 2024 presidential election. This reduced our preregistered index of affective polarization by an average of 0.03 standard deviations (p < 0.05), including a 1.5 degree decrease in differences between the 100 point inparty and outparty feeling thermometers. We saw reductions in active use time for Facebook (-0.37 min/day) and Reddit (-0.2 min/day), but an increase of 0.32 min/day (p < 0.01) for X/Twitter. We saw an increase in reports of negative social media experiences but found no effects on well-being, news knowledge, outgroup empathy, perceptions of and support for partisan violence. This implies that bridging content can improve some societal outcomes without necessarily conflicting with the engagement-driven business model of social media.

cs.SI↗

arXiv2026

Grok in the Wild: Characterizing the Roles and Uses of Large Language Models on Social Media

xAI's large language model, Grok, is called by millions of people each week on the social media platform X. Prior work characterizing how large language models are used has focused on private, one-on-one interactions. Grok's deployment on X represents a major departure from this setting, with interactions occurring in a public social space. In this paper, we systematically sample three months of interaction data to investigate how, when, and to what effect Grok is used on X. At the platform level, we find that Grok responds to 62% of requests, that the majority (51%) are in English, and that engagement is low, with half of Grok's responses receiving 20 or fewer views after 48 hours. We also inductively build a taxonomy of 10 roles that LLMs play in mediating social interactions and use these roles to analyze 41,735 interactions with Grok on X. We find that Grok most often serves as an information provider but, in contrast to LLM use in private one-on-one settings, also takes on roles related to dispute management, such as truth arbiter, advocate, and adversary. Finally, we characterize the population of X users who prompted Grok and find that their self-expressed interests are closely related to the roles the model assumes in the corresponding interactions. Our findings provide an initial quantitative description of human-AI interactions on X, and a broader understanding of the diverse roles that large language models might play in our online social spaces.

cs.SI↗

arXiv2025

Scaling Human Judgment in Community Notes with LLMs

This paper argues for a new paradigm for Community Notes in the LLM era: an open ecosystem where both humans and LLMs can write notes, and the decision of which notes are helpful enough to show remains in the hands of humans. This approach can accelerate the delivery of notes, while maintaining trust and legitimacy through Community Notes' foundational principle: A community of diverse human raters collectively serve as the ultimate evaluator and arbiter of what is helpful. Further, the feedback from this diverse community can be used to improve LLMs' ability to produce accurate, unbiased, broadly helpful notes--what we term Reinforcement Learning from Community Feedback (RLCF). This becomes a two-way street: LLMs serve as an asset to humans--helping deliver context quickly and with minimal effort--while human feedback, in turn, enhances the performance of LLMs. This paper describes how such a system can work, its benefits, key new risks and challenges it introduces, and a research agenda to solve those challenges and realize the potential of this approach.

cs.CY↗

arXiv2025

Community notes reduce engagement with and diffusion of false information online

Social networks scaffold the diffusion of information on social media. Much attention has been given to the spread of true vs. false content on online social platforms, including the structural differences between their diffusion patterns. However, much less is known about how platform interventions on false content alter the engagement with and diffusion of such content. In this work, we estimate the causal effects of Community Notes, a novel fact-checking feature adopted by X (formerly Twitter) to solicit and vet crowd-sourced fact-checking notes for false content. We gather detailed time series data for 40,078 posts for which notes have been proposed and use synthetic control methods to estimate a range of counterfactual outcomes. We find that attaching fact-checking notes significantly reduces the engagement with and diffusion of false content. We estimate that, on average, the notes resulted in reductions of 46.1% in reposts, 44.1% in likes, 21.9% in replies, and 13.5% in views after being attached. Over the posts' entire lifespans, these reductions amount to 11.6% fewer reposts, 13.3% fewer likes, 6.9% fewer replies, and 5.5% fewer views on average. In reducing reposts, we observe that diffusion cascades for fact-checked content are less deep and less "viral," but not less broad, than synthetic control estimates for non-fact-checked content with similar reach. This structural difference contrasts notably with differences between false vs.\ true content diffusion itself, where false information diffuses farther, but with structural patterns that are otherwise indistinguishable from those of true information, conditional on reach.

cs.SI↗

arXiv2024

Reranking partisan animosity in algorithmic social media feeds alters affective polarization

Today, social media platforms hold sole power to study the effects of feed ranking algorithms. We developed a platform-independent method that reranks participants' feeds in real-time and used this method to conduct a preregistered 10-day field experiment with 1,256 participants on X during the 2024 U.S. presidential campaign. Our experiment used a large language model to rerank posts that expressed antidemocratic attitudes and partisan animosity (AAPA). Decreasing or increasing AAPA exposure shifted out-party partisan animosity by two points on a 100-point feeling thermometer, with no detectable differences across party lines, providing causal evidence that exposure to AAPA content alters affective polarization. This work establishes a method to study feed algorithms without requiring platform cooperation, enabling independent evaluation of ranking interventions in naturalistic settings.

cs.CY↗

arXiv2024

Supernotes: Driving Consensus in Crowd-Sourced Fact-Checking

X's Community Notes, a crowd-sourced fact-checking system, allows users to annotate potentially misleading posts. Notes rated as helpful by a diverse set of users are prominently displayed below the original post. While demonstrably effective at reducing misinformation's impact when notes are displayed, there is an opportunity for notes to appear on many more posts: for 91% of posts where at least one note is proposed, no notes ultimately achieve sufficient support from diverse users to be shown on the platform. This motivates the development of Supernotes: AI-generated notes that synthesize information from several existing community notes and are written to foster consensus among a diverse set of users. Our framework uses an LLM to generate many diverse Supernote candidates from existing proposed notes. These candidates are then evaluated by a novel scoring model, trained on millions of historical Community Notes ratings, selecting candidates that are most likely to be rated helpful by a diverse set of users. To test our framework, we ran a human subjects experiment in which we asked participants to compare the Supernotes generated by our framework to the best existing community notes for 100 sample posts. We found that participants rated the Supernotes as significantly more helpful, and when asked to choose between the two, preferred the Supernotes 75.2% of the time. Participants also rated the Supernotes more favorably than the best existing notes on quality, clarity, coverage, context, and argumentativeness. Finally, in a follow-up experiment, we asked participants to compare the Supernotes against LLM-generated summaries and found that the participants rated the Supernotes significantly more helpful, demonstrating that both the LLM-based candidate generation and the consensus-driven scoring play crucial roles in creating notes that effectively build consensus among diverse users.

cs.SI↗

arXiv2024

Reranking Social Media Feeds: A Practical Guide for Field Experiments

Social media plays a central role in shaping public opinion and behavior, yet performing experiments on these platforms and, in particular, on feed algorithms is becoming increasingly challenging. This guide offers practical recommendations for researchers developing and deploying field experiments focused on real-time reranking of social media feeds. The article is organized around two contributions. First, we provide an overview of an experimental method using web browser extensions that intercepts and reranks content in real time, enabling naturalistic reranking field experiments. We then describe feed interventions and measurements that this paradigm enables on participants' actual feeds, without requiring the involvement of social media platforms. Second, we offer concrete technical recommendations for intercepting and reranking social media feeds with minimal user-facing delay, and provide an open-source implementation. This document aims to summarize lessons learned in running field experiments on social media, provide concrete implementation details, and foster the ecosystem of independent social media research. Finally, we release the source code that serves as a blueprint for implementing future feed-ranking experiments.

cs.SI↗

arXiv2024

Off-policy evaluation beyond overlap: partial identification through smoothness

Off-policy evaluation (OPE) is the problem of estimating the value of a target policy using historical data collected under a different logging policy. OPE methods typically assume overlap between the target and logging policy, enabling solutions based on importance weighting and/or imputation. In this work, we approach OPE without assuming either overlap or a well-specified model by considering a strategy based on partial identification under non-parametric assumptions on the conditional mean function, focusing especially on Lipschitz smoothness. Under such smoothness assumptions, we formulate a pair of linear programs whose optimal values upper and lower bound the contributions of the no-overlap region to the off-policy value. We show that these linear programs have a concise closed form solution that can be computed efficiently and that their solutions converge, under the Lipschitz assumption, to the sharp partial identification bounds on the off-policy value. Furthermore, we show that the rate of convergence is minimax optimal, up to log factors. We deploy our methods on two semi-synthetic examples, and obtain informative and valid bounds that are tighter than those possible without smoothness assumptions.

stat.ME↗

arXiv2023

Counterfactual Evaluation of Peer-Review Assignment Policies

Peer review assignment algorithms aim to match research papers to suitable expert reviewers, working to maximize the quality of the resulting reviews. A key challenge in designing effective assignment policies is evaluating how changes to the assignment algorithm map to changes in review quality. In this work, we leverage recently proposed policies that introduce randomness in peer-review assignment--in order to mitigate fraud--as a valuable opportunity to evaluate counterfactual assignment policies. Specifically, we exploit how such randomized assignments provide a positive probability of observing the reviews of many assignment policies of interest. To address challenges in applying standard off-policy evaluation methods, such as violations of positivity, we introduce novel methods for partial identification based on monotonicity and Lipschitz smoothness assumptions for the mapping between reviewer-paper covariates and outcomes. We apply our methods to peer-review data from two computer science venues: the TPDP'21 workshop (95 papers and 35 reviewers) and the AAAI'22 conference (8,450 papers and 3,145 reviewers). We consider estimates of (i) the effect on review quality when changing weights in the assignment algorithm, e.g., weighting reviewers' bids vs. textual similarity (between the review's past papers and the submission), and (ii) the "cost of randomization", capturing the difference in expected quality between the perturbed and unperturbed optimal match. We find that placing higher weight on text similarity results in higher review quality and that introducing randomization in the reviewer-paper assignment only marginally reduces the review quality. Our methods for partial identification may be of independent interest, while our off-policy approach can likely find use evaluating a broad class of algorithmic matching systems.

cs.IR↗

arXiv2021

Engaging Politically Diverse Audiences on Social Media

We study how political polarization is reflected in the social media posts used by media outlets to promote their content online. In particular, we track the Twitter posts of several media outlets over the course of more than three years (566K tweets), and the engagement with these tweets from other users (104M retweets), modeling the relationship between the tweet text and the political diversity of the audience. We build a tool that integrates our model and helps journalists craft tweets that are engaging to a politically diverse audience, guided by the model predictions. To test the real-world impact of the tool, we partner with the PBS documentary series Frontline and run a series of advertising experiments on Twitter. We find that in seven out of the ten experiments, the tweets selected by our model were indeed engaging to a more politically diverse audience, illustrating the effectiveness of our approach.

cs.SI↗

arXiv2021

Perspective-taking to Reduce Affective Polarization on Social Media

The intensification of affective polarization worldwide has raised new questions about how social media platforms might be further fracturing an already-divided public sphere. As opposed to ideological polarization, affective polarization is defined less by divergent policy preferences and more by strong negative emotions towards opposing political groups, and thus arguably poses a formidable threat to rational democratic discourse. We explore if prompting perspective-taking on social media platforms can help enhance empathy between opposing groups as a first step towards reducing affective polarization. Specifically, we deploy a randomized field experiment through a browser extension to 1,611 participants on Twitter, which enables participants to randomly replace their feeds with those belonging to accounts whose political views either agree with or diverge from their own. We find that simply exposing participants to "outgroup" feeds enhances engagement, but not an understanding of why others hold their political views. On the other hand, framing the experience in familiar, empathic terms by prompting participants to recall a disagreement with a friend does not affect engagement, but does increase their ability to understand opposing views. Our findings illustrate how social media platforms might take simple steps that align with business objectives to reduce affective polarization.

cs.SI↗

arXiv2021

Me, My Echo Chamber, and I: Introspection on Social Media Polarization

Homophily -- our tendency to surround ourselves with others who share our perspectives and opinions about the world -- is both a part of human nature and an organizing principle underpinning many of our digital social networks. However, when it comes to politics or culture, homophily can amplify tribal mindsets and produce "echo chambers" that degrade the quality, safety, and diversity of discourse online. While several studies have empirically proven this point, few have explored how making users aware of the extent and nature of their political echo chambers influences their subsequent beliefs and actions. In this paper, we introduce Social Mirror, a social network visualization tool that enables a sample of Twitter users to explore the politically-active parts of their social network. We use Social Mirror to recruit Twitter users with a prior history of political discourse to a randomized experiment where we evaluate the effects of different treatments on participants' i) beliefs about their network connections, ii) the political diversity of who they choose to follow, and iii) the political alignment of the URLs they choose to share. While we see no effects on average political alignment of shared URLs, we find that recommending accounts of the opposite political ideology to follow reduces participants' beliefs in the political homogeneity of their network connections but still enhances their connection diversity one week after treatment. Conversely, participants who enhance their belief in the political homogeneity of their Twitter connections have less diverse network connections 2-3 weeks after treatment. We explore the implications of these disconnects between beliefs and actions on future efforts to promote healthier exchanges in our digital public spheres.

cs.CY↗

arXiv2021

Social Catalysts: Characterizing People Who Spark Conversations Among Others

People assume different and important roles within social networks. Some roles have received extensive study: that of influencers who are well-connected, and that of brokers who bridge unconnected parts of the network. However, very little work has explored another potentially important role, that of creating opportunities for people to interact and facilitating conversation between them. These individuals bring people together and act as social catalysts. In this paper, we test for the presence of social catalysts on the online social network Facebook. We first identify posts that have spurred conversations between the poster's friends and summarize the characteristics of such posts. We then aggregate the number of catalyzed comments at the poster level, as a measure of the individual's "catalystness." The top 1% of such individuals account for 31% of catalyzed interactions, although their network characteristics do not differ markedly from others who post as frequently and have a similar number of friends. By collecting survey data, we also validate the behavioral measure of catalystness: a person is more likely to be nominated as a social catalyst by their friends if their posts prompt discussions between other people more frequently. The measure, along with other conversation-related features, is one of the most predictive of a person being nominated as a catalyst. Although influencers and brokers may have gotten more attention for their network positions, our findings provide converging evidence that another important role exists and is recognized in online social networks.

cs.SI↗

arXiv2021

The Structure of Toxic Conversations on Twitter

Social media platforms promise to enable rich and vibrant conversations online; however, their potential is often hindered by antisocial behaviors. In this paper, we study the relationship between structure and toxicity in conversations on Twitter. We collect 1.18M conversations (58.5M tweets, 4.4M users) prompted by tweets that are posted by or mention major news outlets over one year and candidates who ran in the 2018 US midterm elections over four months. We analyze the conversations at the individual, dyad, and group level. At the individual level, we find that toxicity is spread across many low to moderately toxic users. At the dyad level, we observe that toxic replies are more likely to come from users who do not have any social connection nor share many common friends with the poster. At the group level, we find that toxic conversations tend to have larger, wider, and deeper reply trees, but sparser follow graphs. To test the predictive power of the conversational structure, we consider two prediction tasks. In the first prediction task, we demonstrate that the structural features can be used to predict whether the conversation will become toxic as early as the first ten replies. In the second prediction task, we show that the structural characteristics of the conversation are also predictive of whether the next reply posted by a specific user will be toxic or not. We observe that the structural and linguistic characteristics of the conversations are complementary in both prediction tasks. Our findings inform the design of healthier social media platforms and demonstrate that models based on the structural characteristics of conversations can be used to detect early signs of toxicity and potentially steer conversations in a less toxic direction.

cs.SI↗

arXiv2019

Testing for arbitrary interference on experimentation platforms

Experimentation platforms are essential to modern large technology companies, as they are used to carry out many randomized experiments daily. The classic assumption of no interference among users, under which the outcome of one user does not depend on the treatment assigned to other users, is rarely tenable on such platforms. Here, we introduce an experimental design strategy for testing whether this assumption holds. Our approach is in the spirit of the Durbin-Wu-Hausman test for endogeneity in econometrics, where multiple estimators return the same estimate if and only if the null hypothesis holds. The design that we introduce makes no assumptions on the interference model between units, nor on the network among the units, and has a sharp bound on the variance and an implied analytical bound on the type I error rate. We discuss how to apply the proposed design strategy to large experimentation platforms, and we illustrate it in the context of an experiment on the LinkedIn platform.

stat.ME↗

arXiv2016

Human Atlas: A Tool for Mapping Social Networks

Most social network analyses focus on online social networks. While these networks encode important aspects of our lives they fail to capture many real-world connections. Most of these connections are, in fact, public and known to the members of the community. Mapping them is a task very suitable for crowdsourcing: it is easily broken down in many simple and independent subtasks. Due to the nature of social networks -- presence of highly connected nodes and tightly knit groups -- if we allow users to map their immediate connections and the connections between them, we will need few participants to map most connections within a community. To this end, we built the Human Atlas, a web-based tool for mapping social networks. To test it, we partially mapped the social network of the MIT Media Lab. We ran a user study and invited members of the community to use the tool. In 4.6 man-hours, 22 participants mapped 984 connections within the lab, demonstrating the potential of the tool.

cs.SI↗