Jon Kleinberg
YOU?
Author Swipe
View article: Tracking patterns in toxicity and antisocial behavior over user lifetimes on large social media platforms
Tracking patterns in toxicity and antisocial behavior over user lifetimes on large social media platforms Open
An increasing amount of attention has been devoted to the problem of "toxic" or antisocial behavior on social media. In this paper we analyze such behavior at very large scales: over a 14-year time span on nearly 500 million comments from …
View article: Designing Algorithmic Delegates: The Role of Indistinguishability in Human-AI Handoff
Designing Algorithmic Delegates: The Role of Indistinguishability in Human-AI Handoff Open
As AI technologies improve, people are increasingly willing to delegate tasks to AI agents. In many cases, the human decision-maker chooses whether to delegate to an AI agent based on properties of the specific instance of the decision-mak…
View article: Replicating Electoral Success
Replicating Electoral Success Open
A core tension in the study of plurality elections is the clash between the classic Hotelling-Downs model, which predicts that two office-seeking candidates should cater to the median voter, and the empirical observation that democracies o…
View article: Private Blotto: Viewpoint Competition with Polarized Agents
Private Blotto: Viewpoint Competition with Polarized Agents Open
Social media platforms are responsible for collecting and disseminating vast quantities of content. Recently, however, they have also begun enlisting users in helping annotate this content - for example, to provide context or label disinfo…
View article: The Backfiring Effect of Weak AI Safety Regulation
The Backfiring Effect of Weak AI Safety Regulation Open
Recent policy proposals aim to improve the safety of general-purpose AI, but there is little understanding of the efficacy of different regulatory approaches to AI safety. We present a strategic model that explores the interactions between…
View article: AI-Assisted Decision Making with Human Learning
AI-Assisted Decision Making with Human Learning Open
AI systems increasingly support human decision-making. In many cases, despite the algorithm's superior performance, the final decision remains in human hands. For example, an AI may assist doctors in determining which diagnostic tests to r…
View article: Maia-2: A Unified Model for Human-AI Alignment in Chess
Maia-2: A Unified Model for Human-AI Alignment in Chess Open
There are an increasing number of domains in which artificial intelligence (AI) systems both surpass human ability and accurately model human behavior. This introduces the possibility of algorithmically-informed teaching in these domains t…
View article: What's in a Niche? Migration Patterns in Online Communities
What's in a Niche? Migration Patterns in Online Communities Open
Broad topics in online platforms represent a type of meso-scale between individual user-defined communities and the whole platform; they typically consist of related communities that address different facets of a shared topic. Users often …
View article: Tracking Patterns in Toxicity and Antisocial Behavior Over User Lifetimes on Large Social Media Platforms
Tracking Patterns in Toxicity and Antisocial Behavior Over User Lifetimes on Large Social Media Platforms Open
An increasing amount of attention has been devoted to the problem of "toxic" or antisocial behavior on social media. In this paper we analyze such behavior at very large scales: we analyze toxicity over a 14-year time span on nearly 500 mi…
View article: A comprehensive generalization of the Friendship Paradox to weights and attributes
A comprehensive generalization of the Friendship Paradox to weights and attributes Open
The Friendship Paradox is a simple and powerful statement about node degrees in a graph. However, it only applies to undirected graphs with no edge weights, and the only node characteristic it concerns is degree. Since many social networks…
View article: Evaluating the World Model Implicit in a Generative Model
Evaluating the World Model Implicit in a Generative Model Open
Recent work suggests that large language models may implicitly learn world models. How should we assess this possibility? We formalize this question for the case where the underlying reality is governed by a deterministic finite automaton.…
View article: How Random is Random? Evaluating the Randomness and Humaness of LLMs' Coin Flips
How Random is Random? Evaluating the Randomness and Humaness of LLMs' Coin Flips Open
One uniquely human trait is our inability to be random. We see and produce patterns where there should not be any and we do so in a predictable way. LLMs are supplied with human data and prone to human biases. In this work, we explore how …
View article: Designing Skill-Compatible AI: Methodologies and Frameworks in Chess
Designing Skill-Compatible AI: Methodologies and Frameworks in Chess Open
Powerful artificial intelligence systems are often used in settings where they must interact with agents that are computationally much weaker, for example when they work alongside humans or operate in complex environments where some tasks …
View article: Content Moderation and the Formation of Online Communities: A Theoretical Framework
Content Moderation and the Formation of Online Communities: A Theoretical Framework Open
We study the impact of content moderation policies in online communities. In our theoretical model, a platform chooses a content moderation policy and individuals choose whether or not to participate in the community according to the fract…
View article: Language Generation in the Limit
Language Generation in the Limit Open
Although current large language models are complex, the most basic specifications of the underlying language generation problem itself are simple to state: given a finite set of training samples from an unknown language, produce valid new …
View article: Arbitrariness and Social Prediction: The Confounding Role of Variance in Fair Classification
Arbitrariness and Social Prediction: The Confounding Role of Variance in Fair Classification Open
Variance in predictions across different trained models is a significant, under-explored source of error in fair binary classification. In practice, the variance on some data examples is so large that decisions can be effectively arbitrary…
View article: The Moderating Effect of Instant Runoff Voting
The Moderating Effect of Instant Runoff Voting Open
Instant runoff voting (IRV) has recently gained popularity as an alternative to plurality voting for political elections, with advocates claiming a range of advantages, including that it produces more moderate winners than plurality and co…
View article: On the Actionability of Outcome Prediction
On the Actionability of Outcome Prediction Open
Predicting future outcomes is a prevalent application of machine learning in social impact domains. Examples range from predicting student success in education to predicting disease risk in healthcare. Practitioners recognize that the ulti…
View article: Modeling reputation-based behavioral biases in school choice
Modeling reputation-based behavioral biases in school choice Open
A fundamental component in the theoretical school choice literature is the problem a student faces in deciding which schools to apply to. Recent models have considered a set of schools of different selectiveness and a student who is unsure…
View article: Replicating Electoral Success
Replicating Electoral Success Open
A core tension in the study of plurality elections is the clash between the classic Hotelling-Downs model, which predicts that two office-seeking candidates should position themselves at the median voter's policy, and the empirical observa…
View article: Equilibria, Efficiency, and Inequality in Network Formation for Hiring and Opportunity
Equilibria, Efficiency, and Inequality in Network Formation for Hiring and Opportunity Open
Professional networks -- the social networks among people in a given line of work -- can serve as a conduit for job prospects and other opportunities. Here we propose a model for the formation of such networks and the transfer of opportuni…
View article: Microstructures and Accuracy of Graph Recall by Large Language Models
Microstructures and Accuracy of Graph Recall by Large Language Models Open
Graphs data is crucial for many applications, and much of it exists in the relations described in textual format. As a result, being able to accurately recall and encode a graph described in earlier text is a basic yet pivotal ability that…
View article: From Graphs to Hypergraphs: Hypergraph Projection and its Remediation
From Graphs to Hypergraphs: Hypergraph Projection and its Remediation Open
We study the implications of the modeling choice to use a graph, instead of a hypergraph, to represent real-world interconnected systems whose constituent relationships are of higher order by nature. Such a modeling choice typically involv…
View article: Hypergraph patterns and collaboration structure
Hypergraph patterns and collaboration structure Open
Humans collaborate in different contexts such as in creative or scientific projects, in workplaces and in sports. Depending on the project and external circumstances, a newly formed collaboration may include people that have collaborated b…
View article: Using large language models to promote health equity
Using large language models to promote health equity Open
Advances in large language models (LLMs) have driven an explosion of interest about their societal impacts. Much of the discourse around how they will impact social equity has been cautionary or negative, focusing on questions like "how mi…
View article: Informational Diversity and Affinity Bias in Team Growth Dynamics
Informational Diversity and Affinity Bias in Team Growth Dynamics Open
Prior work has provided strong evidence that, within organizational settings, teams that bring a diversity of information and perspectives to a task are more effective than teams that do not. If this form of informational diversity confers…