Chenhao Tan (谭宸浩)

I am an associate professor at the Department of Computer Science and Data Science at the University of Chicago. I direct the Chicago Human+AI lab (CHAI) (, ). You can learn more about my life trajectory and find my official biography here.

My research aims to develop the best AI for humans. I am mainly interested in the following problems.

• AI & Scientific Discovery. We aim to 1) develop research agents to improve scientific research at all stages (e.g., idea explorer and hypothesis generation); 2) work with domain scientists (check out AI and Science online seminar); and 3) investigate theories on scientist-AI collaboration, both on incentives and tacit knowledge.
• Understanding and improving LLMs. The most direct application of AI in science for me is to AI itself. We aim to understand model behavior through both prompting science and mechanistic interpretability (e.g., concept incongruence and MechEvalAgents). I am also interested in novel ways that LLMs can unlock human understanding (e.g., communication games).
• AI & Democratic Processes. AI is reshaping how we can produce and consume information. In addition to scientific discovery, I am interested in how AI will or can influence democratic processes. We are actively working on CivicChat (stay tuned).

Driven by exciting progress in AI, I find myself updating this description at much higher pace than in the past. If you are interested in seeing how my research description has evolved, you can check this page.

Weekly fun:

• We are running weekly Agents4Science competition, where you vote on ideas and we will implement them with research agents for you! Please vote!
• Guess whether Ari or I wrote an AI hot take: Ari or Chenhao!

News:

• I moved my website to be hosted on Github recently due to a server issue. Some old files are not recovered yet. If you need them, send me an email and I will let you know when they are back.
• New blog: The Mirage of Autonomous AI Scientists.
• I am co-organizing an online seminar on AI & Scientific Discovery. Check it out!
• We have developed a communication game called HR Simulator™. See if you can handle office drama better than LLMs!

Selected recent work:

• Haokun Liu*, Yangqiaoyu Zhou*, Mingxuan Li*, Chenfei Yuan, and Chenhao Tan. Literature Meets Data: A Synergistic Approach to Hypothesis Generation. [Website]
• Yangqiaoyu Zhou, Haokun Liu, Tejes Srivastava, Hongyuan Mei, and Chenhao Tan. Hypothesis Generation with Large Language Models. NLP4Science workshop at EMNLP 2024.
• Ari Holtzman and Chenhao Tan. Prompting as Scientific Inquiry. NeurIPS 2025 (position papers track). [Blog]
• [Xiaoyan Bai][xiaoyan_bai], Ike Peng, Aditya Singh, and Chenhao Tan. Concept Incongruence: An Exploration of Time and Death in Role Playing. NeurIPS 2025. [Blog]

Recent talk:

Openings:

As of fall 2025, you can apply to work with me through

• the UChicago Computer Science PhD program,
• the UChicago Data Science PhD program,
• the Data Science Instituite Postdoctoral program (If you applied to the DSI postdoc program and mentioned me as an advisor, please send me an email because I can potentially support you with my own funding).

I also have a postdoc opening on novel intelligence that can start any time. Please email me your CV and names of references. I am always looking for motivated masters and undergraduate students who are interested in Human-centered AI, Communication & Intelligence, AI & Scientific Discovery, and AI alignment. Please read this FAQ before contacting me.

Demos & Packages

• Hypogenic AI
• HR Simulator
• Hypothesis generation
• Task delegability, should AI be used for X?

If you want to learn more about my research, check my papers. And kudos to all my collaborators!