Raymond Fok

Raymond Fok is a human-computer interaction researcher who builds and studies AI tools for working with scientific papers, and he co-authored a NeurIPS 2025 Best Paper on how alike language models' answers are. At the University of Michigan, he worked in the Crowds + Machines Lab on a UIST 2017 paper about crowd-powered email search and was a 2018 finalist for CRA's Outstanding Undergraduate Researcher Award. He co-authored a CHI 2021 study of AI explanations and team performance, then wrote Scim, a skimming interface that began as an internship at Semantic Scholar while he was a University of Washington PhD student and appeared at IUI 2023. Qlarify, with its recursively expandable abstracts, followed in October 2023, and Scideator, a human-LLM system for facet-based ideation, in September 2024. In 2025, he co-authored Artificial Hivemind, whose Infinity-Chat dataset showed that different language models produce strikingly similar outputs, and as of October 2026 he is affiliated with Microsoft.

Now at: Microsoft
Badges: Bridge

Timeline

2017

Crowd-powered email search

As a computer science undergraduate at the University of Michigan, he works in the Crowds + Machines (CroMa) Lab and helps build a privacy-preserving, crowd-powered email search system with researchers at Carnegie Mellon's Human Computer Interaction Institute. His extractor generation algorithm turns crowd-generated examples into regular expressions that run against a user's emails, and the resulting paper is accepted to UIST 2017.
2018

A CRA research award finalist

In 2018, he is named a finalist for CRA's Outstanding Undergraduate Researcher Award. He works with Walter S. Lasecki on human-computer interaction systems for non-traditional computers such as smart watches and virtual assistants.
2021

AI explanations and team performance

In 2021, he is a coauthor, with Gagan Bansal, Tongshuang Wu, Joyce Zhou, Besmira Nushi, Ece Kamar, Marco Tulio Ribeiro and Daniel S. Weld, of a CHI 2021 study of AI explanations. In user studies on three datasets, AI augmentation produces complementary improvements that explanations leave unchanged, and explanations raise the chance that people accept the AI's recommendation regardless of its correctness.
2023

The Scim skimming interface

In 2023, he is first author of Scim, an intelligent skimming interface that highlights paper content faceted by content type, with a density readers can set. He builds the first version as an internship project at Semantic Scholar while a PhD student at the University of Washington, and the paper appears at IUI 2023.
October 2023

Qlarify and expandable abstracts

In October 2023, he posts Qlarify with Joseph Chee Chang, Tal August, Amy X. Zhang and Daniel S. Weld. The system introduces recursively expandable abstracts, which use retrieval-augmented generation to extend an abstract with information from the paper's full text and link each addition to its source passages.
September 2024

Scideator for scientific ideation

In September 2024, he coauthors Scideator with Marissa Radensky, Simra Shahid, Pao Siangliulue, Tom Hope and Daniel S. Weld. The human-LLM system extracts purposes, mechanisms and evaluations from papers so that users can recombine them into new ideas.
2025

A NeurIPS 2025 Best Paper

In 2025, he coauthors Artificial Hivemind with Liwei Jiang, Yejin Choi and others, and NeurIPS names it one of four best papers. The Infinity-Chat dataset in the paper shows that a single language model repeats itself and that different models produce strikingly similar outputs.
As of October 2026

Affiliated with Microsoft

As of October 2026, he is affiliated with Microsoft, with human-AI interaction and NLP as his research interests.