ReDDDoT Phase 1: Planning Grant: Writing Education through Design-Oriented AI

Billions of people rely on software to prepare written communication. These software are now equipped with chatbot-style interfaces that automate tasks of text production without adequate safeguards. These safeguards aim to sustain the wide-ranging proficiencies that go into effective communication or maintain the quality and integrity of research and written work. This project will lay the groundwork for new paradigms of responsible design, development, and deployment of artificial intelligence (AI)-powered writing tools with a focus on the “college writing” domain at a large top-tier university. The project’s interdisciplinary collaborations will connect instructors and learners of college writing, researchers in natural language processing (NLP), and domain experts from the humanities, to set an agenda for research in generative AI with the potential to enhance rich practices of research, writing, and communication in and beyond higher education. The design-oriented approach supports and empowers community-led participation in technology design and keeps the public interest firmly in view. The project combines outreach to instructors and students of "writing" at Rutgers University (a broad category based on core writing requirements) and design labs that put insights gained from this outreach in dialogue with the goals and methods of AI research. Outreach efforts will clarify: 1) the learning objectives that writing communities aspire to and achieve; 2) the strengths, weaknesses, values, and aspirations that they bring to these endeavors; 3) the current capabilities that writing tools already offer, including observed strengths and limitations; and 4) a technically specific and socially rich account of the needs and expectations for building community-centered paradigms for responsible AI. Design labs will enable the documentation of meaningful ways that these communities can leverage existing and near-term AI tools in their writing practices–with the right kinds of preparation and facilitation—and will lead to new perspectives on AI infrastructures and architectures that could significantly streamline future efforts but might require longer-term research and investment. The vision for the outcome of this planning project includes: 1) clear and accessible articulations of the research methods and challenges for design-oriented AI as it applies to writing; 2) models for the equitable and sustainable trajectory of the technology in the college classroom; and 3) resources that make it easier to pursue research in this area, such as software infrastructure, preliminary results, hypotheses, open questions, and an inclusive, collaborative research network. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.