Research Overview

CARE conducts cutting-edge research across seven key directions to ensure AI systems are designed and used safely, reliably, and beneficially for humanity.

Seven Key Research Directions

Our interdisciplinary approach brings together experts from various fields to address the complex challenges of creating safe and responsible AI systems.

AI Safety + Science

Developing safety protocols and evaluation methods for AI systems used in scientific research, with a focus on preventing misuse while enabling beneficial applications in chemistry, biology, and materials science.

Partners: Lawrence Livermore National Lab, Center for AI Safety
Learn more →

AI Safety + Data/Computing

Creating infrastructure for secure data collection, efficient model training, and thorough evaluation of AI systems, ensuring that computational resources are used responsibly and safely across the AI development pipeline.

Partners: National Center for Supercomputing Applications, Cloud Providers
Learn more →

AI Safety + Algorithms/Theoretical Foundations

Advancing the theoretical foundations of AI safety through novel algorithms, formal verification methods, and robust optimization techniques that provide guarantees about AI system behavior and performance.

Partners: UIUC Computer Science Department, MIT CSAIL
Learn more →

AI Safety + Medicine/Healthcare

Creating robust safety mechanisms for AI systems deployed in healthcare settings, ensuring they provide reliable, unbiased recommendations while maintaining patient privacy and adhering to medical ethics.

Partners: UIUC Medical School, Healthcare Industry Partners
Learn more →

AI Safety + Economics

Analyzing the economic implications of widespread AI deployment, with a focus on developing sustainable economic models that promote innovation while protecting workers and creative professionals.

Partners: Berkeley HAAS School of Business, Economic Policy Institute
Learn more →

AI Safety + Law/Policy

Developing policy frameworks and regulatory approaches that balance innovation with safety, addressing questions of liability, accountability, and governance for increasingly capable AI systems.

Partners: Stanford Center for Legal Informatics, Georgetown Law
Learn more →

AI Safety + Social Good

Harnessing AI for positive social impact while ensuring safety and equity, focusing on applications that address global challenges like climate change, public health, and education access.

Partners: UN AI for Good, Partnership on AI
Learn more →

Recent Publications

Our researchers are actively publishing their findings in top-tier academic conferences and journals. Explore our recent publications below.

MMDT: Decoding the Trustworthiness and Safety of Multimodal Foundation Models

Chejian Xu, Jiawei Zhang, Zhaorun Chen, Chulin Xie, Mintong Kang, Yujin Potter, Zhun Wang, Zhuowen Yuan, Alexander Xiong, Zidi Xiong, Chenhui Zhang, Lingzhi Yuan, Yi Zeng, Peiyang Xu, Chengquan Guo, Andy Zhou, Jeffrey Ziwei Tan, Xuandong Zhao, Francesco Pinto, Zhen Xiang, Yu Gai, Zinan Lin, Dan Hendrycks, Bo Li, Dawn Song
ICLR 2025

AIR-Bench 2024: A Safety Benchmark Based on Risk Categories from Regulations and Policies

Yi Zeng, Yu Yang, Andy Zhou, Jeffrey Ziwei Tan, Yuheng Tu, Yifan Mai, Kevin Klyman, Minzhou Pan, Ruoxi Jia, Dawn Song, Percy Liang, Bo Li
ICLR 2025

DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models

Boxin Wang, Weixin Chen, Hengzhi Pei, Chulin Xie, Mintong Kang, Chenhui Zhang, Chejian Xu, Zidi Xiong, Ritik Dutta, Rylan Schaeffer, Sang T. Truong, Simran Arora, Mantas Mazeika, Dan Hendrycks, Zinan Lin, Yu Cheng, Sanmi Koyejo, Dawn Song, Bo Li.
NeurIPS 2023
View All Publications