We study AI’s role in writing, editing, and summarizing information.
Meaning and representation
In medical notes, news, and research, wording matters. We ask how AI changes meaning, reliability, and whose perspectives appear.
Why it matters
A summary determines what a reader sees first and what they may never see. When institutions use AI to write or edit, small changes in wording can affect how people, events, and uncertainty are represented. Understanding those changes matters to the people producing a record, the people described in it, and those who later rely on it.
Questions we ask
- What does AI preserve, omit, or change when it summarizes a person's account?
- How does AI assistance affect variation in writing and the perspectives that appear?
- How can people evaluate a rewritten text in relation to its original purpose and audience?
Our approach
We use natural language processing and social research to examine language in context. Comparing an original account with a generated version can make changes visible, but judging those changes also requires understanding the setting. Accuracy, readability, voice, and representation may each call for different forms of evaluation.
Support this research
Support could make room for students to study human and AI writing, develop comparisons that capture meaningful changes, and document methods for others to use. A focused collaboration can define the kind of text, the people affected, and the evidence needed to assess a proposed use of AI.
Discuss supporting this researchUpdated