Do frontier LLMs represent how national cultures change?
Evaluated Claude, GPT, Gemini and Qwen against two decades of World Values Survey data across 40 countries.
PhD Student · Social Psychology · USC · Expected 2028
I study how human values change across time and cultures and what LLMs are doing to them.
Quantitative behavioral scientist working on AI evaluation, trust in AI, and multilingual & cultural safety. My focus is measurement: operationalizing constructs like trust, bias, harm and cultural alignment so evaluations capture what they claim to.
◆ Open to Summer 2027 research internships
Research themes
01
Culture, Language & AI
Whether LLMs represent people outside the Western, English-speaking world, how they are reshaping cultures, and where safety fails in other languages.
02
Trust & reliance on AI
How people come to trust AI systems, when that trust is warranted, and how relying on them changes decisions.
03
AI, identity & society
The sociotechnical and ethical consequences of living with these systems, including how interacting with AI shapes people's sense of self.
04
Values & moral judgement
How moral values shift over time and differ across societies, and how groups misread each other's values (the psychology the AI work is built on).
Selected work
All research →The Homogenizing Engine: AI's Role in Standardizing Culture and the Path to Policy
Examining how AI systems may standardize cultural representations and what this means for policy and cultural diversity.
Asymmetric Moral Metaperceptions: Consequences for Political Polarization and Paths to Correction
10M+ tweets and three preregistered experiments show a low-cost correction raises intergroup trust.
In the media
USC Dornsife News
AI is changing more than your writing — it may be shaping your worldview
Coverage of "The Homogenizing Engine" on how AI systems can flatten cultural differences, and what policy can do about it.
Read the story →
Tech Xplore · April 2026
AI is changing more than your writing—it may be shaping your worldview, say researchers
Coverage of "The Homogenizing Engine" and its findings on whose worldview AI systems reproduce.
Read the story →
SPSP Newsletter
Yalda Daryani develops field guide for science advocacy
SPSP featured and published the science advocacy field guide I wrote after the AAAS CASE Workshop.
Read the story →Experience
Full experience →
June 2026 — August 2026
Google Public Policy Fellow
— Center for Democracy & Technology
LLM Evals · Multilingual AI Safety · Sovereign AI
February 2026 — July 2026
Trustworthy AI Researcher
— Digital Trust Council
AI Governance & Certification
April 2026 — Present
Graduate Researcher
— USC Mind & Society Center
Value change · AI & Identity
August 2023 — April 2026
Graduate Researcher
— USC Morality & Language Lab
AI & Culture · Moral Psychology
News
2026
First-author preprint on LLMs and national cultural change
is now on arXiv.
Summer 2026
Joined the Center for Democracy & Technology as a
Google Public Policy Fellow.
Feb 2026
Presented moral misperceptions research at SPSP.
Dec 2025
"The Homogenizing Engine" published.
Recognition
Graduate Student Poster Award,
SPSP · 2025
Dornsife Ph.D. Academy Scholarship & Research Fund
· 2024, 2025
Dornsife/Graduate School Fellowship,
USC · 2023
Outstanding Research Award,
SPSP · 2021
+ 4 departmental grants and travel funds (see CV)