Hello! I'm an AI Safety Generalist specializing in fieldbuilding at the intersection between AI safety and the social sciences. Specifically, I focus on leveraging the deep, underutilized talent pool of social science researchers for AI safety.
My own research focus within that sits in empirical and technical AI governance, spanning two threads: (1) how AI systems replicate human behavior and the attendant consequences for society, with particular attention to how these systems pick up and reproduce human biases, personas, and social dynamics; and (2) how organizations approach and adopt AI systems as an emerging technology, with particular attention to how they identify, communicate about, and manage the risks that come with that adoption.
I’m currently a Research Mentor at SPAR, where I lead projects investigating the above topics. Previously, I earned an MA in Media Studies / Sociology from Columbia University under the advisement of Diane Vaughan and James Chu. Before that, I studied Communication and Political Science at Brigham Young University-Hawaii, where I got my start in research with Mason Allred and read a lot of dead philosophers. I also spent a summer as a graduate researcher in the Management Division at Columbia Business School with Alan Zhang.
During my time at Columbia, I served as a Research Manager and AI Governance Fellowship Facilitator with the Columbia AI Alignment Club (CAIAC). In the former, my research group investigated how language models exhibited human behaviors, including truth-bias, ideological bias, and escalation of commitment. In the latter, I led ~25 students through the fundamentals of the field and guided them in generating new research ideas in AI Safety.
In terms of personal interests, I spend lots of time and energy on running, swimming, reading, music (jazz, classical, indie), philosophy, photography, and many other things. Previously, I volunteered in the Republic of Kiribati for nine months; paced 45 miles of the HURT 100; and backpacked through Switzerland, Austria, and Italy with my wife.
Feel free to reach out for any reason; I’m always happy to chat with new people.
Publications
Preprints
-
Normalization of Deviance in AI Development
Emilio Barkett, Daniel Graham, Alexander Kimpton
NeurIPS 2026 Trustworthy AI for Good (underreview)
-
Whose Alignment? Process Alignment Across Organizational Contexts
Niklas Weller, Emilio Barkett
ICML 2026 Pluralistic Alignment Workshop
-
Representation Without Control: Testing the Realization Effect in Language Models
Ciarán Walsh, Emilio Barkett
NeurIPS 2026 SocialAgent Workshop (underreview)
-
The Compulsory Imaginary: AGI and Corporate Authority
Emilio Barkett
Preprint
-
Status Hierarchies in Language Models
Emilio Barkett
Master's Thesis / Preprint
-
Getting out of the Big-Muddy: Escalation of Commitment in LLMs
Emilio Barkett, Olivia Long, Paul Kröger
NeurIPS 2026 SocialAgent Workshop (underreview)
-
Don’t Change My View: Ideological Bias Auditing in Large Language Models
Paul Kröger, Emilio Barkett
NeurIPS 2026 SocialAgent Workshop (underreview)
-
Reasoning Isn’t Enough: Examining Truth-Bias and Sycophancy in LLMs
Emilio Barkett, Olivia Long, and Madhavendra Thakur
ICML 2025 Workshop on Models of Human Feedback for AI Alignment (MoFA)
Works in Progress
-
Legislative Boundary-Work and the Professional Response to Artificial Intelligence
Emilio Barkett, [looking for co-authors]
Working Paper
-
The Model Compass: Linguistic Suppression as an Indicator of Democratic Backsliding
Lauren Benjamin Mushro, Shruti Das, Emilio Barkett
Working Paper
-
Evaluating Moral Reasoning in Embodied Language Model Agents
Roshni Lulla, Kristin Witte, Saloni Modi, Emilio Barkett
Working Paper (SPAR)
-
Underrepresented Languages in LLMs
Raghav Pant, Dan Sachs, Emilio Barkett
Working Paper (SPAR)
-
Illusion of Explanatory Depth in LLMs
Hayzie Chu, Andrew Xie, Emilio Barkett
Working Paper (SPAR)
Contributions as an RA
-
Peace from the Pulpit? Religious Appeals Do Not Outperform Secular Appeals in Reducing Partisan Animosity among American Christians
James Chu, Yejin Park Roberts
Under Review