Hello! My name is Nari (rhymes with "starry") and I am a fourth-year PhD student in Carnegie Mellon University's Machine Learning Department, where I'm advised by Hoda Heidari.

I do interdisciplinary research that explores how practitioners can govern, anticipate, measure, and mitigate risks posed by AI systems. My work draws from theories and methods in AI/ML, HCI, design, and policy research to study the societal impacts of AI. These days, I am especially interested in methods that center the lived experiences of people who might be negatively impacted by AI, and interventions that give these stakeholders a say in how AI is developed and governed.

Previously, I graduated from Harvard in 2021 with a BA and MS in Computer Science, where I worked with Finale Doshi-Velez. I've also interned at Microsoft Research UK on the Teachable AI Experiences Team.

Email: narij at andrew dot cmu dot edu

Selected Research

For a complete list, see my Google Scholar.

Public Procurement for Responsible AI? Understanding US Cities' Practices, Challenges, and Needs
Nari Johnson, Elise Silva, Harrison Leon, Motahhare Eslami, Beth Schwanke, Ravit Dotan, Hoda Heidari
Oral Presentation at the NeurIPS Regulatable ML Workshop
arXiv preprint, under revision
October 2024 | pdf| abstract| cite| response to US OMB RFI|

The Fall of an Algorithm: Characterizing the Dynamics Toward Abandonment
Nari Johnson, Sanika Moharana, Christina N. Harrington, Nazanin Andalibi, Hoda Heidari*, Motahhare Eslami*
ACM Conference on Fairness, Accountability, and Transparency (FAccT), 2024
(* = Equal Contribution)
June 2024 | pdf| abstract| database| podcast episode| cite|

Where Does My Model Underperform? A Human Evaluation of Slice Discovery Algorithms
Nari Johnson, Ángel Alexander Cabrera, Gregory Plumb, Ameet Talwalkar
🏆 HCOMP 2023 Best Paper Award (top 1)
Spotlight Presentation (top 5%) at the ICML SCIS Workshop
AAAI Conference on Human Computation and Crowdsourcing (HCOMP), 2023
ICML Workshop on Spurious Correlations, Invariance, and Stability, 2023
November 2023 | pdf| abstract| cite| conference talk| data & code|

Provocation: Who benefits from "inclusion" in Generative AI?
Nari Johnson*, Samantha Dalal*, Siobhan Mackenzie Hall*
Oral Presentation at the NeurIPS EvalEval Workshop
The NeurIPS 2024 Workshop on Evaluating Evaluations
(* = Equal Contribution, shared first authorship)
October 2024 | pdf| abstract| cite|