Back
Emily Bender - Faculty Homepage (University of Washington Linguistics)
webfaculty.washington.edu·faculty.washington.edu/ebender/
Emily Bender is a key critic of LLM scaling hype and a prominent researcher linking linguistics with AI safety concerns; her 'Stochastic Parrots' paper sparked major debate in the AI community about responsible development.
Metadata
Importance: 62/100homepage
Summary
Homepage of Emily Bender, Professor of Linguistics at the University of Washington, known for influential work on language model risks, including co-authoring the 'Stochastic Parrots' paper and formulating the 'Octopus Test' and Bender Rule. Her research examines harms from large language models, AI ethics, and the gap between linguistic form and meaning.
Key Points
- •Co-authored 'On the Dangers of Stochastic Parrots' (2021), a landmark paper critiquing the uncritical scaling of large language models
- •Developed the 'Bender Rule' advocating that NLP papers always name the language(s) their systems are trained/evaluated on
- •Research focuses on meaning, form, and the limitations of LLMs in understanding language vs. mimicking it
- •Prominent voice on AI harms, ethics, and responsible NLP research practices
- •Engages in public scholarship and policy discussions around AI safety and governance
Review
Emily Bender is a prominent computational linguist who has made significant contributions to understanding the ethical and societal implications of language technologies, particularly large language models and AI systems. Her research centers on grammar engineering, linguistic typology, and critically examining the potential harms of computational language technologies. Bender's work is distinguished by her interdisciplinary approach, combining deep linguistic expertise with critical analysis of technology's social impacts. She has been a leading voice in highlighting the potential risks of AI systems, particularly large language models, and advocating for more responsible and ethically-aware development of natural language processing technologies. Her research spans grammar engineering, sociolinguistic variation, and the broader societal consequences of computational language technologies.
Cached Content Preview
HTTP 200Fetched Apr 4, 202610 KB
Emily M. Bender
Emily M. Bender
Professor, Director,
Professional MS in Computational Linguistics
Preferred contact method
email
Social media
Mastodon : @emilymbender@dair-community.social
Twitter : @emilymbender
Office hours/office location
Email for info
About Me
I have a been a member of the faculty at the University of
Washington since 2003. I am currently the Thomas L. and Margo G. Wyckoff Endowed Professor in the Department of
Linguistics and the faculty director of the CLMS program and the director
of the Computational
Linguistics Laboratory . For 2019-2022, I was honored to be the
Howard and Frances Nostrand Endowed Professor. I am an Adjunct
Professor in both the School of
Computer Science and Engineering and the Information School at UW, and a
member of the Tech Policy
Lab , Value Sensitive Design
Lab , and RAISE .
I am the past Chair (2016-2017) of the Executive
Board of NAACL and have previously
served as a member of the ICCL (2014-2018;
the committee responsible for Coling ). I
am serving on the Executive Board of the Association
for Computational Linguistics from 2022-2025 as VP Elect, VP,
President and then Past President. I am a fellow of the AAAS (American Association for the Advancement of Science; 2022).
Prior
to coming to UW, I held temporary positions at Stanford University and
UC Berkeley, and worked in industry at YY Technologies. I received my
PhD from the Linguistics Department at Stanford University, where I
joined the HPSG and LinGO projects at CSLI . My AB (also in
Linguistics) is from UC Berkeley , and I've also studied at Tohoku
University in Sendai, Japan.
In 2012, LINGUIST List asked me to write
an essay about how I came to be a linguist and in 2018, the LSA interviewed me for their member spotlight feature. My pronouns are she/her and my Erdős
number is 4. My Bacon-Erdős number, allowing director links, is 7.
Research Interests
Multilingual Grammar Engineering
My grammar engineering work centers on the LinGO Grammar Matrix , an
open-source starter kit for the development of broad-coverage
precision HPSG grammars. These grammars map strings to detailed
linguistic representations in the framework of Minimal Recursion
Semantics. The AGGREGATION
project is investigating the automatic creation of grammars from IGT
with the Grammar Matrix for the benefit of language documentation.
The Grammar Matrix is developed in the context of the DELPH-IN consortium , and
Matrix-derived grammars are compatible with the DELPH-IN suite of
open-source tools. The Grammar Matrix itself represents an approach
to computational linguistic typology, using computational methodology
to combine depth of formal methods (creating grammars which map
surface strings to
... (truncated, 10 KB total)Resource ID:
10202ee006b2ebdf | Stable ID: sid_7ki9aEl9G2