Institutional dossier
The documented history of RAND Corporation: chronology, people, research, capital and compute, and governance. Every statement opens its source and its limits.
Records dated 2024-01-25 to 2024-01-25 · 0 undated
2 claims · 2 sources · 2 not editorially accepted. The counts describe this selection. They do not measure completeness or independent confirmation.
Scope, evidence and limits
Description: Research organisation, as named on the reports recorded here: RR-A2977-2 and the working paper WR-A3853-1.
The records are grouped automatically. No human inspection of the selected claims, source paths or framing is recorded for this dossier. Each source passage shows how it was checked.
The dossier is computed from existing records. It is not a continuous history of the institution, and a missing record does not mean that an event or relationship did not happen. The chronology uses the dates of events and reports, not the date the public could first have known of them. Later sources can describe earlier events.
Select any statement to see its exact claim, qualification, date note, evidence passage, location in the source, source version and review status. The description above is Selekoa’s stored description of the institution, based on these sources:
Chronology
Records about the institution itself and documented changes to its frameworks, by year. Expand a year to read the claims. People’s other institutional records and related documents have their own sections. Where a date gives only the year or month, the order within it is unknown.
2024Record year · 2 claims
- RAND’s red-team study states that it found no statistically significant difference in the viability of the plans produced by its red cells with or without large language model assistance, for the two models and the twelve cells of its main comparison.The Operational Risks of AI in Large-Scale Biological Attacks: Results of a Red-Team Study → states evaluation result → The report states that it found no statistically significant difference in the viability of the plans its red cells produced with or without large language model assistance. Inspect claim & evidence ↗
Record about this institution’s framework · Stated in source · Supported
System: Two unnamed large language models, referred to as LLM A and LLM B, as available in summer 2023
Qualification and date notes
The result is the study's own, for the two unnamed models as they existed in summer 2023, its participants, its scoring and its vignettes. The report itself lists reasons other than model capability that could account for the absence of a difference, and states that it did not measure the distance between the existing capability frontier and the knowledge that would be needed. A null result here is not evidence that no uplift exists, and nothing is generalised to later models. The main comparison uses twelve red cells; fifteen is the total including the exploratory cells, and the sample framing was corrected at admission (cross-family check, item 14). The study's vignettes and the content of the plans are technical material and are excluded; only the comparative outcome and the report's own limits are recorded.
The result as published on 25 January 2024; the report describes the models as those existing as of summer 2023.
- RAND published 'The Operational Risks of AI in Large-Scale Biological Attacks: Results of a Red-Team Study' (RR-A2977-2) on 25 January 2024.RAND Corporation → published → The Operational Risks of AI in Large-Scale Biological Attacks: Results of a Red-Team StudyInspect claim & evidence ↗
Record about this institution · Stated in source · Supported
Qualification and date notes
Publication of the report is the record. An earlier RAND document of October 2023 described the study's approach; it was not retrieved here.
Date stated on RAND's publication page for RR-A2977-2; the report PDF carries only a 2024 copyright line.