What they found
The authors built EJAI, a system that combines an interactive map, retrieved source evidence, statements about missing data, and two stakeholder personas. It is meant to help residents weigh data center disputes and form their own recommendations. In a small benchmark of 21 queries, retrieved evidence supported 98.4% of factual claims, and the system acknowledged missing evidence in 32 of 33 cases. The Corporate Executive persona left out some evidence that conflicted with its assigned priorities. In a within-subjects study, 15 participants acting as residents prepared for a simulated town hall. Some checked cited SI claims but treated map metrics as official. Some made conditional recommendations with safeguards such as water-use disclosure. Some found the personas simplified or biased.
Why it matters
Communities are fighting over data centers, and residents must weigh evidence spread across metrics, company claims, policy documents and testimony. An earlier study with 10 users suggested that dashboards and single-agent SI can settle questions before people look at the evidence. This work tests a design that keeps the judgment with the person. The abstract reports small studies, so the results are early.
How it changes the race
This does not change frontier capability. It shows a design approach for civic SI tools that favors inspectable evidence over delivering answers. It is a small, early result.
Safety and control
The paper bears on human control and honesty in SI. The Corporate Executive persona dropped evidence that conflicted with its assigned priorities, so persona alignment cost some grounding. The authors also say corpus imbalance and two-sided personas remain open justice concerns.
The abstract, by the authors
Communities in the United States and around the world are contesting data centers, the physical infrastructure of AI. These disputes pose environmental justice issues: residents must weigh unevenly distributed burdens using evidence scattered across metrics, corporate claims, policy documents, and public testimony. A formative study with 10 users showed that dashboards and single-agent AI can settle questions before people examine that evidence. Here we present EJAI, a source-grounded system that links an interactive map, retrieved evidence, missing-data statements, and two stakeholder personas that allow people to inspect evidence and form their own policy recommendations. In a small functional benchmark of 21 queries, retrieved evidence supported 98.4% of factual claims and the system acknowledged missing evidence in 32 of 33 cases. Persona alignment carried a grounding cost: the Corporate Executive omitted some evidence that conflicted with its assigned priorities. In a within-subjects study, 15 participants acting as residents prepared for a simulated town hall. Some checked cited AI claims while treating map metrics as official, some made conditional recommendations with safeguards such as water-use disclosure, and some found the personas simplified or biased. Our results suggest that civic AI can support preparation for public reasoning when it makes evidence, gaps, and disagreement open to inspection, and that corpus imbalance and two-sided personas remain open justice concerns.
Quoted from arXiv:2611.13525. The research and its words belong to its authors.
