Blog
Personal perspectives on AI research, trustworthy AI, AI for science, and teaching. For research write-ups, see the research pages and the DREAM Lab blog.
-
Will AI Replace Scientists? Human–AI Collaboration Wins Only When the Human Adds Something
Studies show that people working with AI often do worse than AI alone. In open-ended science, the scientist still adds what AI lacks: the question, the context, and the check.
-
Why Alzheimer's Patients With the Same Diagnosis Differ
One diagnosis of Alzheimer's disease covers patients whose brains, proteins, and genes follow different paths. We should study each patient's profile, not only the average patient.
-
Sounding Like You Know
Frontier models are getting better at sounding right. In biology, where answers are slow to check, that fluency is harder to catch than a plain error.
-
Math Is Not Just Problem Solving
AI can now write research-level math proofs. Our new study asks whether models also track the hidden hypotheses that learning math trains us to notice.
-
Vision Should Be Structural
Why my group spent a few years turning images into SVG code that language models can read, and what I think still holds.
-
A Unified View of Trustworthy AI
Robustness, fairness, and interpretability are usually studied apart, but many of their methods answer the same question about data.
-
We Evaluate Prompt-Injection Defenses Too Simply
A low attack success rate can mean a model handled untrusted text correctly, or that it threw the text away, and the standard metric does not say which.