A Design Space for Intelligent and Interactive Writing Assistants

Welcome to our design space for intelligent and interactive writing assistants! The design space consists of five aspects: task, user, technology, interaction, and ecosystem. Within each aspect, we define dimensions (i.e., fundamental components of an aspect) and codes (i.e., potential options for each dimension). Please refer to our paper for the detailed definitions of each dimension and code.

To create this design space, we collaborated with researchers from a variety of disciplines, including Human-Computer Interaction (HCI), Natural Language Processing (NLP), Information Systems, and Education, and annotated 115 papers from HCI and NLP fields to understand the current landscape of writing assistants. We hope that our design space offers researchers and designers a practical tool to navigate, comprehend, and compare the various possibilities of writing assistants, and aid in the envisioning and design of new writing assistants.

Our design space is a living artifact, as it will evolve over time alongside the fields. We invite the community to contribute to this artifact by adding new papers, annotations, and discussions to track future developments in this space.

$The figure shows the integrated design space for writing assistants. It is comprised of five boxes, Ecosystem, Task, User, Interaction, and Technology. \n Ecosystem is the biggest box that surrounds the rest, and has the following items: Digital infrastructure (e.g., usability consistency, technical interoperability), Social factors (e.g., designing with stakeholders, desigining for social writing), Locale (e.g., local writing, remote writing), Access model (e.g., free and/or open-source software, commercial software), Norms and Rules (e.g., laws, conventions), and Change over time (e.g., authors, readers, writing, information environment). \n The top part inside Ecosystem box is occupied by Task box, which overlaps with User, Interaction, and Technology boxes. Task box has the following items: Writing stage (e.g., planning, drafting, revision), Writing context (e.g., academic, journalistic, technical), Purpose (e.g., expository, narrative, descriptive), Specificity (e.g., general direction, detailed requirements), and Audience (e.g., specified, implied).\n Interaction box is positioned in the middle of the User and Technology boxes, with arrows connecting to and from these boxes. Interaction has the following points: User - Steering the system (e.g., explicit, implicit, no control), User - Integrating system output (e.g., selection, inspiration), UI - Interface paradigm (e.g., text editor, chatbot), UI - Layout (e.g., writing area, separated, input UI), UI - Visual differentiation (e.g., formatting, location), UI - Interaction metaphor (e.g., agent, tool, hybrid), UI - Initiation (e.g., user-initiated, system-initiated), Technology - Output type (e.g., analysis, generation), Technology - Curation type (e.g., deterministic, curated options), and Technology - User Data access (e.g., input text, additional data). \n User box is on the left of the Interaction box, with the following points: Demographic profile (e.g., age, language and culture), User capabilities (e.g., writing expertise, efficiency), Relationship to system (e.g., agency, ownership, trust), and System output preferences (e.g., coherence, diversity). \n Technology box is on the left of Interaction box and below the User box, with the following points: Data - Source (e.g., experts, users), Data - Size (e.g., small, medium, large), Model - Type (e.g., rule-based, foundation models), Model - External resource access (e.g., tool, data), Learning - Problem (e.g., classification, generation), Learning - Algorithm (e.g., supervised, unsupervised), Learning - Training and adaptation (e.g., fine-tuning, prompting), Evaluation - Evaluator (e.g., automatic, machine-learned), Evaluation - Focus (e.g., linguistic quality, controllability), and Scalability (e.g., cost, latency).$

Want to add your paper/writing assistant to the list? Please either (i) fill out this Google form or (ii) create a pull request in our GitHub repository
Have questions or found incorrect annotation? Please email Mina Lee
Interested in contributing to the project? Please visit our GitHub repository and start contributing!

Annotated Papers

Best used with a larger screen.


2024	Drafting, Revision	Technical, Professional, Personal	Descriptive, Persuasive, Entertainment	Specific Objectives	Specified	Profession	Profession			Agency, Ownership, Trust, Transparency	Agency, Ownership, Trust, Transparency					Foundation model	Tool				Human evaluation	Linguistic quality, Style & adequacy	Latency	Explicit	Selection	Text editor	Writing area	Formatting, Location	Tool	User-initiated, System-initiated	Proposal	Model	Input text	Usability consistency		Free and open-source software
2023	Revision	Professional	Feedback, Persuasive	Detailed Requirements, General Direction	Specified			Textual coherence	Textual coherence	Ownership, Availability, Trust	Ownership, Availability, Trust, Agency	Writing expertise	Writing expertise	Expert	Unknown	Statistical ML model, Rule-based model	Data	Regression	Supervised learning	Fine-tuning	Machine-learned evaluation	Linguistic quality	Cost	Implicit, Explicit	Inspiration, No integration	Other, Text editor	Separated	Location	Tool	User-initiated	Analysis	Deterministic	Input text		Designing with stakeholders			Writing
2023														Machine, Expert	Large (<1M)	Foundation model	Data	Generation	Supervised learning	Fine-tuning	Automatic evaluation, Human evaluation	Linguistic quality	Cost
2023	Revision	Professional, Academic	Feedback, Persuasive	General Direction, Specific Objectives	Implied	Education	Education	Textual coherence	Textual coherence	Ownership	Ownership	Writing expertise	Writing expertise	User, Machine	Large (<1M)	Deep neural network	Data	Classification	Supervised learning	Fine-tuning	Automatic evaluation	Linguistic quality	Cost
2023	Drafting	Other	Descriptive	Specific Objectives	Specified	Other, Profession	Other, Profession	Textual diversity, Explainability	Textual coherence, Explainability	Other	Other	Efficiency	Efficiency			Rule-based model	Tool				Human evaluation	Style & adequacy		No control	Inspiration, No integration	Text editor	Separated	Location	Tool	System-initiated	Proposal, Analysis	Curated	Input text		Designing with stakeholders			Writing
2023	Revision	Professional	Educational, Persuasive	Detailed Requirements, Specific Objectives	Specified	Education	Education	Other, Explainability	Other, Explainability	Ownership, Trust	Ownership, Trust	Writing expertise, Other	Writing expertise, Other, Efficiency	Expert	Large (<1M)	Foundation model	Data	Classification	Supervised learning	Fine-tuning	Human evaluation, Automatic evaluation	Other		Implicit	Inspiration	Text editor	Separated	Location, Media type, Formatting	Tool	User-initiated	Analysis	Model	Input text		Designing with stakeholders		Conventions	Writing
2023						Profession		Personalization		Ownership	Availability, Trust	Writing expertise																							Designing with stakeholders		Conventions
2023	Revision		Feedback	Detailed Requirements			Language & culture	Textual coherence	Textual coherence			Writing expertise	Writing expertise	Expert, Other, User	Extremely large (>1M)	Foundation model		Rewriting, Classification	Supervised learning	Fine-tuning	Automatic evaluation, Human-machine evaluation	Linguistic quality, Style & adequacy									Analysis	Model	Input text
2023						Language & culture, Profession		Other		Other		Efficiency		Authors	Medium (<10k)
2023	Revision	Academic	Feedback, Persuasive	Detailed Requirements, Specific Objectives	Implied	Education	Education	Other				Writing expertise, Other	Writing expertise, Other	User, Expert	Medium (<10k)	Foundation model, Deep neural network		Classification	Supervised learning	Fine-tuning	Automatic evaluation, Human evaluation, Human-machine evaluation	Other

0 of 116 row(s) selected.

Rows per page

Page 1 of 12

Authors: Mina Lee , Katy Ilonka Gero, John Joon Young Chung, Simon Buckingham Shum, Vipul Raheja, Hua Shen, Subhashini Venugopalan, Thiemo Wambsganss, David Zhou, Emad A. Alghamdi, Tal August, Avinash Bhat, Madiha Zahrah Choksi, Senjuti Dutta, Jin L.C. Guo, Md Naimul Hoque, Yewon Kim, Seyed Parsa Neshaei, Agnia Sergeyuk, Antonette Shibani, Disha Shrivastava, Lila Shroff, Jessi Stark, Sarah Sterman, Sitong Wang, Antoine Bosselut, Daniel Buschek, Joseph Chee Chang, Sherol Chen, Max Kreminski, Joonsuk Park, Roy Pea, Eugenia H. Rho, Shannon Zejiang Shen, Pao Siangliulue

Core group of annotators: Avinash Bhat, Simon Buckingham Shum, Agnia Sergeyuk, Yewon Kim, David Zhou, Emad A. Alghamdi, Jin L.C. Guo, Seyed Parsa Neshaei, Hua Shen, Md Naimul Hoque, Madiha Zahrah Choksi, Katy Ilonka Gero, Sarah Sterman, Antonette Shibani, Mina Lee

Designer of this artifact: Shannon Zejiang Shen, Mina Lee

A Design Space for Intelligent and Interactive Writing Assistants

Annotated Papers

Beyond the Chat: Executable and Verifiable Text-Editing with LLMs

Your Text Is Hard to Read: Facilitating Readability Awareness to Support Writing Proficiency in Text Production

Visualize Before You Write: Imagination-Guided Open-Ended Text Generation

To Revise or Not to Revise: Learning to Detect Improvable Claims for Argumentative Writing Support

Supporting Novices Author Audio Descriptions via Automatic Feedback

Structured Persuasive Writing Support in Legal Education: A Model and Tool for German Legal Case Solutions

Social Dynamics of AI Support in Creative Writing

Smart Word Suggestions for Writing Assistance

Semi-Automated Approach for Evaluating Severe Weather Risk Communication

Predicting Desirable Revisions of Evidence and Reasoning in Argumentative Writing