Post, November 13, 2025 • (National Institute of Standards and Technology) AgentDojo-Inspect is a codebase created by the U.S. AI Safety Institute to facilitate research into agent hijacking and defenses against said hijacking. Agent hijacking is a type of indirect prompt injection [1] in which an attacker inserts malicious instructions into data that may be ingested...
Human-AI Teaming
Current Filters
Clear all
Data on the effects of transparency on people’s perceptions of social AI
Post, November 13, 2025 • (University of Michigan) The data was collected from a survey study using Qualtrics. In this between-subject experimental vignette study, we aimed to investigate the impact of introducing a chatbot with varying levels of transparency and different framings (intelligent entity vs. machine) on participants' perceptions of social chatbots. We presented participants...
Decision Making Under Uncertainty Human Subjects Data – Fire Evacuation Task
Post, November 13, 2025 • (Sandia National Laboratories) This dataset contains de-identified data from human subjects experiments, along with the images and code that were used to run the experiments (as a crowdsourced online study). In this study, participants were shown the probability of a house being in the burn zone of a wildfire. They...
Decision Making Under Uncertainty Human Subjects Data – Wind Speed Forecasts
Post, November 13, 2025 • (Sandia National Laboratories) This dataset contains de-identified data from three human subjects experiments (SNL000464) in which participants completed a mock public safety power shutoff task. The participants were instructed to turn off a circuit if they thought the wind speed would go higher than 50 mph in a certain location....
Experimental data on trial-by-trial AI assistance, feedback, human confidence and decisions during an AI-assisted decision-making chess puzzle task
Post, November 13, 2025 • (Carnegie Mellon University, others) The data are collected from a human subjects experiment to examine how human confidence in AI and self-confidence change and impact their decision-making during an AI-assisted decision-making task. The experimental task includes three practice and 30 experimental chess puzzle problems, which 100 participants are asked to...
Human-AI Interactions Dataset
Post, November 13, 2025 • (Stanford University) This repo provides access to our dataset of over 30,000 Human-AI interactions. We recruited crowdworkers and expert dermatologists and asked them to perform a variety of binary classification tasks. Participants gave two responses -- their initial impression after seeing the task, and their response after being presented with...
Meta-Analysis of Performance in Human-AI Systems
Post, November 13, 2025 • (Massachusetts Institute of Technology) When combinations of humans and AI are useful: A systematic review and meta-analysis Link to Dataset Meta-Analysis of Performance in Human-AI SystemsDate last updated: July 10, 2023
Travian Dataset
Post, November 13, 2025 • (Sandia National Laboratories, others) The aim of this project is to study the structure of different types of networks that emerge in massively multi-player online games. Travian is a popular browser-based real-time strategy game with more than 5 million players. Players form alliances of up to 60 members under a...