Reliable information · Science communication · Sustainable AI
Dustin Wright
Tenure-Track Assistant Professor
I study how AI changes information: whether generated and summarised content remains traceable to evidence, faithful, and diverse; whether its evaluation remains robust; and whether the systems producing it are environmentally and socially sustainable.
Dustin develops methods for reliable information across the path from source evidence to generated and public-facing text. His work spans scientific claim generation and fact-checking, information change in science communication, evidence-attributed long-context summarisation, uncertainty under distribution shift, and stress tests of factuality metrics. Recent work broadens this programme to the values and epistemic diversity of language-model outputs, alongside collaborative work on whether language models faithfully simulate human dialogue.
He joined AAU NLP as a Tenure-Track Assistant Professor in February 2026. Previously, he held a Danish Data Science Academy postdoctoral fellowship at the University of Copenhagen, including a research visit to the University of Michigan, and worked in UCPH’s SAINTS Lab. Across technical work on efficient modelling and a systems perspective on environmental sustainability, he argues that computational efficiency is one part—not the definition—of sustainable AI.
Evidence and reliable information
Factuality, faithfulness, attribution, uncertainty, and evaluation that remains valid under controlled change and distribution shift.
Science communication and diversity
Fact-checking, information change, public-facing science, model values, and the diversity of claims made accessible through AI.
Sustainable AI
Combining efficient modelling with systems-level analysis of environmental and social consequences.
Selected work
5 representative papers
Unstructured Evidence Attribution for Long Context Query Focused Summarization
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Efficiency Is Not Enough: A Critical Perspective on Environmentally Sustainable AI
- Communities, Domains & Consequences
BMRS: Bayesian Model Reduction for Structured Pruning
- Communities, Domains & Consequences
LLM Tropes: Revealing Fine-Grained Values and Opinions in Large Language Models
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
Generating Scientific Claims for Zero-Shot Scientific Fact Checking
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Supervision and collaboration
Questions worth working on
I am interested in projects on factuality, evidence attribution, science communication, information diversity, robust evaluation, efficient modelling, and technically grounded questions about AI sustainability.