Emily M. Bender
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Emily M. Bender has a total of 52 co-authors. Here are the top co-authors by number of shared publications:
- Olga Zamaraeva
University of Washington
3 shared publications - Antske Fokkens
Computational Linguistics and Text Mining Lab Vrije Universiteit Amsterdam
2 shared publications - Fei Xia
University of Washington
2 shared publications - Kristen Howell
LivePerson Inc University of Washington
2 shared publications Laurie Poulson
University of Washington
2 shared publications- Michael Wayne Goodman
Nanyang Technological University Singapore
2 shared publications
Latest Publications
- Top Cited Papers
- 2 February 2026
- Published by University of Illinois Libraries in First Monday
De-anthropomorphizing “AI”: From wishful mnemonics to accurate nomenclature
Language matters. How we describe “AI” technology influences how it is perceived, deployed, and trusted. Extravagant and persuasive language incites hype. It is the responsibility of journalists, companies, and scholars to characterize technology in ways that inform and empower their readers by using appropriate terminology and avoiding inflated claims. One type of inflated claim comes from using anthropomorphizing language to describe system functionality. Anthropomorphization is the attribution of human capabilities and characteristics to the inanimate system. In this paper, we present a linguistic analysis of anthropomorphizing language in 29 texts (a total of 1,368 sentences) from academic articles, online news articles, and company blog posts. We construct a taxonomy of eight categories of anthropomorphization: Cognizer, Products of cognition, Emotion, Communication, Agent, Human role analogy, Names and pronouns, and Biological metaphors. Following this taxonomy we present concrete strategies for how to de-anthropomorphize the language we use to describe “AI” based on a functionality-first principle.
- 2 February 2026
- Published by University of Illinois Libraries in First Monday
Top Cited PapersDe-anthropomorphizing “AI”: From wishful mnemonics to accurate nomenclature
Language matters. How we describe “AI” technology influences how it is perceived, deployed, and trusted. Extravagant and persuasive language incites hype. It is the responsibility of journalists, companies, and scholars to characterize technology in ways that inform and empower their readers by using appropriate terminology and avoiding inflated claims. One type of inflated claim comes from using anthropomorphizing language to describe system functionality. Anthropomorphization is the attribution of human capabilities and characteristics to the inanimate system. In this paper, we present a linguistic analysis of anthropomorphizing language in 29 texts (a total of 1,368 sentences) from academic articles, online news articles, and company blog posts. We construct a taxonomy of eight categories of anthropomorphization: Cognizer, Products of cognition, Emotion, Communication, Agent, Human role analogy, Names and pronouns, and Biological metaphors. Following this taxonomy we present concrete strategies for how to de-anthropomorphize the language we use to describe “AI” based on a functionality-first principle.
- 26 August 2025
- Published by Ubiquity Press, Ltd. in Journal of Interactive Media in Education
- Vol. 2025 (1)
Unsafe AI for Education: A Conversation on Stochastic Parrots and Other Learning Metaphors ⚠
- 26 August 2025
- Published by Ubiquity Press, Ltd. in Journal of Interactive Media in Education
- Vol. 2025 (1)
Unsafe AI for Education: A Conversation on Stochastic Parrots and Other Learning Metaphors ⚠
- Top Cited PapersOpen Access
- 27 January 2025
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences of the United States of America
- Vol. 122 (5)
How should the advancement of large language models affect the practice of science?
Large language models (LLMs) are being increasingly incorporated into scientific workflows. However, we have yet to fully grasp the implications of this integration. How should the advancement of large language models affect the practice of science? For this opinion piece, we have invited four diverse groups of scientists to reflect on this query, sharing their perspectives and engaging in debate. Schulz et al. make the argument that working with LLMs is not fundamentally different from working with human collaborators, while Bender et al. argue that LLMs are often misused and overhyped, and that their limitations warrant a focus on more specialized, easily interpretable tools. Marelli et al. emphasize the importance of transparent attribution and responsible use of LLMs. Finally, Botvinick and Gershman advocate that humans should retain responsibility for determining the scientific roadmap. To facilitate the discussion, the four perspectives are complemented with a response from each group. By putting these different perspectives in conversation, we aim to bring attention to important considerations within the academic community regarding the adoption of LLMs and their impact on both current and future scientific practices.
- 27 January 2025
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences of the United States of America
- Vol. 122 (5)
Top Cited PapersOpen AccessHow should the advancement of large language models affect the practice of science?
Large language models (LLMs) are being increasingly incorporated into scientific workflows. However, we have yet to fully grasp the implications of this integration. How should the advancement of large language models affect the practice of science? For this opinion piece, we have invited four diverse groups of scientists to reflect on this query, sharing their perspectives and engaging in debate. Schulz et al. make the argument that working with LLMs is not fundamentally different from working with human collaborators, while Bender et al. argue that LLMs are often misused and overhyped, and that their limitations warrant a focus on more specialized, easily interpretable tools. Marelli et al. emphasize the importance of transparent attribution and responsible use of LLMs. Finally, Botvinick and Gershman advocate that humans should retain responsibility for determining the scientific roadmap. To facilitate the discussion, the four perspectives are complemented with a response from each group. By putting these different perspectives in conversation, we aim to bring attention to important considerations within the academic community regarding the adoption of LLMs and their impact on both current and future scientific practices.
- Open Access
- 26 November 2024
- Published by Not Crossref prefixes - transferred records in Proceedings of the International Conference on Head-Driven Phrase Structure Grammar
- p. 76–87
The morphological causative in Panãra: A Grammar Matrix implementation
In this paper, we describe the modeling of the causative morpheme in an implemented HPSG grammar for the Panãra language. The Panãra causative morpheme appears on the verb between the agreement markers for the ergative and absolutive arguments, increasing the verb's valency to two core arguments. Due to the linear ordering of inflectional morphemes (position classes) in Grammar Matrix grammars, the position of the causative morpheme would require the verb to have information about its valence before the morpheme is attached. We model the descriptive data with a new lexical rule that introduces the causative morpheme and changes the verbal inflection path from intransitive to transitive using the existing Valence Change library. This analysis is validated in an implemented grammar fragment for Panãra.
- 26 November 2024
- Published by Not Crossref prefixes - transferred records in Proceedings of the International Conference on Head-Driven Phrase Structure Grammar
- p. 76–87
Open AccessThe morphological causative in Panãra: A Grammar Matrix implementation
In this paper, we describe the modeling of the causative morpheme in an implemented HPSG grammar for the Panãra language. The Panãra causative morpheme appears on the verb between the agreement markers for the ergative and absolutive arguments, increasing the verb's valency to two core arguments. Due to the linear ordering of inflectional morphemes (position classes) in Grammar Matrix grammars, the position of the causative morpheme would require the verb to have information about its valence before the morpheme is attached. We model the descriptive data with a new lexical rule that introduces the causative morpheme and changes the verbal inflection path from intransitive to transitive using the existing Valence Change library. This analysis is validated in an implemented grammar fragment for Panãra.
- Open Access
- 3 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 2322-2347
From "AI" to Probabilistic Automation: How Does Anthropomorphization of Technical Systems Descriptions Influence Trust?
In this paper we investigate how people’s level of trust (as reported through self-assessment) in so-called “AI” (artificial intelligence) is influenced by anthropomorphizing language in system descriptions. Building on prior work, we define four categories of anthropomorphization (1. Properties of a cognizer, 2. Agency, 3. Biological metaphors, and 4. Properties of a communicator). We use a survey-based approach (n=954) to investigate whether participants are likely to trust one of two (fictitious) “AI” systems by randomly assigning people to see either an anthropomorphized or a de-anthropomorphized description of the systems. We find that participants are no more likely to trust anthropomorphized over de-anthropmorphized product descriptions overall. The type of product or system in combination with different anthropomorphic categories appears to exert greater influence on trust than anthropomorphizing language alone, and age is the only demographic factor that significantly correlates with people’s preference for anthropomorphized or de-anthropomorphized descriptions. When elaborating on their choices, participants highlight factors such as lesser of two evils, lower or higher stakes contexts, and human favoritism as driving motivations when choosing between product A and B, irrespective of whether they saw an anthropomorphized or a de-anthropomorphized description of the product. Our results suggest that “anthropomorphism” in “AI” descriptions is an aggregate concept that may influence different groups differently, and provide nuance to the discussion of whether anthropomorphization leads to higher trust and over-reliance by the general public in systems sold as “AI”.
- 3 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 2322-2347
Open AccessFrom "AI" to Probabilistic Automation: How Does Anthropomorphization of Technical Systems Descriptions Influence Trust?
In this paper we investigate how people’s level of trust (as reported through self-assessment) in so-called “AI” (artificial intelligence) is influenced by anthropomorphizing language in system descriptions. Building on prior work, we define four categories of anthropomorphization (1. Properties of a cognizer, 2. Agency, 3. Biological metaphors, and 4. Properties of a communicator). We use a survey-based approach (n=954) to investigate whether participants are likely to trust one of two (fictitious) “AI” systems by randomly assigning people to see either an anthropomorphized or a de-anthropomorphized description of the systems. We find that participants are no more likely to trust anthropomorphized over de-anthropmorphized product descriptions overall. The type of product or system in combination with different anthropomorphic categories appears to exert greater influence on trust than anthropomorphizing language alone, and age is the only demographic factor that significantly correlates with people’s preference for anthropomorphized or de-anthropomorphized descriptions. When elaborating on their choices, participants highlight factors such as lesser of two evils, lower or higher stakes contexts, and human favoritism as driving motivations when choosing between product A and B, irrespective of whether they saw an anthropomorphized or a de-anthropomorphized description of the product. Our results suggest that “anthropomorphism” in “AI” descriptions is an aggregate concept that may influence different groups differently, and provide nuance to the discussion of whether anthropomorphization leads to higher trust and over-reliance by the general public in systems sold as “AI”.
- Top Cited PapersOpen Access
- 15 April 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on the Web
- Vol. 18 (3) , 1-24
Envisioning Information Access Systems: What Makes for Good Tools and a Healthy Web?
We observe a recent trend toward applying large language models (LLMs) in search and positioning them as effective information access systems. While the interfaces may look appealing and the apparent breadth of applicability is exciting, we are concerned that the field is rushing ahead with a technology without sufficient study of the uses it is meant to serve, how it would be used, and what its use would mean. We argue that it is important to reassert the central research focus of the field of information retrieval, because information access is not merely an application to be solved by the so-called ‘AI’ techniques du jour. Rather, it is a key human activity, with impacts on both individuals and society. As information scientists, we should be asking what do people and society want and need from information access systems and how do we design and build systems to meet those needs? With that goal, in this conceptual article we investigate fundamental questions concerning information access from user and societal viewpoints. We revisit foundational work related to information behavior, information seeking, information retrieval, information filtering, and information access to resurface what we know about these fundamental questions and what may be missing. We then provide our conceptual framing about how we could fill this gap, focusing on methods as well as experimental and evaluation frameworks. We consider the Web as an information ecosystem and explore the ways in which synthetic media, produced by LLMs and otherwise, endangers that ecosystem. The primary goal of this conceptual article is to shed light on what we still do not know about the potential impacts of LLM-based information access systems, how to advance our understanding of user behaviors, and where the next generations of students, scholars, and developers could fruitfully invest their energies.
- 15 April 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on the Web
- Vol. 18 (3) , 1-24
Top Cited PapersOpen AccessEnvisioning Information Access Systems: What Makes for Good Tools and a Healthy Web?
We observe a recent trend toward applying large language models (LLMs) in search and positioning them as effective information access systems. While the interfaces may look appealing and the apparent breadth of applicability is exciting, we are concerned that the field is rushing ahead with a technology without sufficient study of the uses it is meant to serve, how it would be used, and what its use would mean. We argue that it is important to reassert the central research focus of the field of information retrieval, because information access is not merely an application to be solved by the so-called ‘AI’ techniques du jour. Rather, it is a key human activity, with impacts on both individuals and society. As information scientists, we should be asking what do people and society want and need from information access systems and how do we design and build systems to meet those needs? With that goal, in this conceptual article we investigate fundamental questions concerning information access from user and societal viewpoints. We revisit foundational work related to information behavior, information seeking, information retrieval, information filtering, and information access to resurface what we know about these fundamental questions and what may be missing. We then provide our conceptual framing about how we could fill this gap, focusing on methods as well as experimental and evaluation frameworks. We consider the Web as an information ecosystem and explore the ways in which synthetic media, produced by LLMs and otherwise, endangers that ecosystem. The primary goal of this conceptual article is to shed light on what we still do not know about the potential impacts of LLM-based information access systems, how to advance our understanding of user behaviors, and where the next generations of students, scholars, and developers could fruitfully invest their energies.
- Open Access
- 20 March 2024
- Published by Association for Computing Machinery (ACM) in ACM Journal on Responsible Computing
- Vol. 1 (1) , 1-17
Data Statements: From Technical Concept to Community Practice
Responsible computing ultimately requires that technical communities develop and adopt tools, processes, and practices that mitigate harms and support human flourishing. Prior efforts toward the responsible development and use of datasets, machine learning models, and other technical systems have led to the creation of documentation toolkits to facilitate transparency, diagnosis, and inclusion. This work takes the next step: to catalyze community uptake, alongside toolkit improvement. Specifically, starting from one such proposed toolkit specialized for language datasets, data statements for natural language processing, we explore how to improve the toolkit in three senses: (1) the content of the toolkit itself, (2) engagement with professional practice, and (3) moving from a conceptual proposal to a tested schema that the intended community of use may readily adopt. To achieve these goals, we first conducted a workshop with natural language processing practitioners to identify gaps and limitations of the toolkit as well as to develop best practices for writing data statements, yielding an interim improved toolkit. Then we conducted an analytic comparison between the interim toolkit and another documentation toolkit, datasheets for datasets. Based on these two integrated processes, we present our revised Version 2 schema and best practices in a guide for writing data statements. Our findings more generally provide integrated processes for co-evolving both technology and practice to address ethical concerns within situated technical communities.
- 20 March 2024
- Published by Association for Computing Machinery (ACM) in ACM Journal on Responsible Computing
- Vol. 1 (1) , 1-17
Open AccessData Statements: From Technical Concept to Community Practice
Responsible computing ultimately requires that technical communities develop and adopt tools, processes, and practices that mitigate harms and support human flourishing. Prior efforts toward the responsible development and use of datasets, machine learning models, and other technical systems have led to the creation of documentation toolkits to facilitate transparency, diagnosis, and inclusion. This work takes the next step: to catalyze community uptake, alongside toolkit improvement. Specifically, starting from one such proposed toolkit specialized for language datasets, data statements for natural language processing, we explore how to improve the toolkit in three senses: (1) the content of the toolkit itself, (2) engagement with professional practice, and (3) moving from a conceptual proposal to a tested schema that the intended community of use may readily adopt. To achieve these goals, we first conducted a workshop with natural language processing practitioners to identify gaps and limitations of the toolkit as well as to develop best practices for writing data statements, yielding an interim improved toolkit. Then we conducted an analytic comparison between the interim toolkit and another documentation toolkit, datasheets for datasets. Based on these two integrated processes, we present our revised Version 2 schema and best practices in a guide for writing data statements. Our findings more generally provide integrated processes for co-evolving both technology and practice to address ethical concerns within situated technical communities.
- Top Cited Papers
- 2 February 2024
- Published by SAGE Publications in Current Directions in Psychological Science
- Vol. 33 (2) , 114-120
Resisting Dehumanization in the Age of “AI”
The production and promotion of “AI” technology involves dehumanization on many fronts. I explore these processes of dehumanization and the role that cognitive science can play by bringing a richer picture of human cognition to the discourse.
- 2 February 2024
- Published by SAGE Publications in Current Directions in Psychological Science
- Vol. 33 (2) , 114-120
Top Cited PapersResisting Dehumanization in the Age of “AI”
The production and promotion of “AI” technology involves dehumanization on many fronts. I explore these processes of dehumanization and the role that cognitive science can play by bringing a richer picture of human cognition to the discourse.
- Open Access
Neural Text Generation from Rich Semantic Representations
We propose neural models to generate high-quality text from structured representations based on Minimal Recursion Semantics (MRS). MRS is a rich semantic representation that encodes more precise semantic detail than other representations such as Abstract Meaning Representation (AMR). We show that a sequence-to-sequence model that maps a linearization of Dependency MRS, a graph-based representation of MRS, to English text can achieve a BLEU score of 66.11 when trained on gold data. The performance can be improved further using a high-precision, broad coverage grammar-based parser to generate a large silver training corpus, achieving a final BLEU score of 77.17 on the full test set, and 83.37 on the subset of test data most closely matching the silver data domain. Our results suggest that MRS-based representations are a good choice for applications that need both structured semantics and the ability to produce natural language text as output.
Open AccessNeural Text Generation from Rich Semantic Representations
We propose neural models to generate high-quality text from structured representations based on Minimal Recursion Semantics (MRS). MRS is a rich semantic representation that encodes more precise semantic detail than other representations such as Abstract Meaning Representation (AMR). We show that a sequence-to-sequence model that maps a linearization of Dependency MRS, a graph-based representation of MRS, to English text can achieve a BLEU score of 66.11 when trained on gold data. The performance can be improved further using a high-precision, broad coverage grammar-based parser to generate a large silver training corpus, achieving a final BLEU score of 77.17 on the full test set, and 83.37 on the subset of test data most closely matching the silver data domain. Our results suggest that MRS-based representations are a good choice for applications that need both structured semantics and the ability to produce natural language text as output.
- 15 October 2018
- Published by Not Crossref prefixes - transferred records in Proceedings of the International Conference on Head-Driven Phrase Structure Grammar
- p. 66–86
Nominalized clauses in the Grammar Matrix
We present an analysis of clausal nominalization developed in the context of the LinGO Grammar Matrix (Bender et al., 2002, 2010) to support the addition of subordinate clauses to the grammar customization framework. In particular, we examine the typological variation of nominalized clausal complements and nominalized clausal modifiers. To account for the range of variation in nominalized clauses across the world's languages and to support linguists in exploring alternative analyses, we propose a flexible library of analyses, allowing nominalization of the clause to occur at the V, VP or S level.
- 15 October 2018
- Published by Not Crossref prefixes - transferred records in Proceedings of the International Conference on Head-Driven Phrase Structure Grammar
- p. 66–86
Nominalized clauses in the Grammar Matrix
We present an analysis of clausal nominalization developed in the context of the LinGO Grammar Matrix (Bender et al., 2002, 2010) to support the addition of subordinate clauses to the grammar customization framework. In particular, we examine the typological variation of nominalized clausal complements and nominalized clausal modifiers. To account for the range of variation in nominalized clauses across the world's languages and to support linguists in exploring alternative analyses, we propose a flexible library of analyses, allowing nominalization of the clause to occur at the V, VP or S level.