Discovery · side-by-side comparison
USDA and DOJ
Similarity 0.56 (strong). Strong: descriptions share much specific vocabulary; usually the same kind of task. Classified as Potentially reusable capability: DOJ reports this as deployed while USDA reports pre-deployment. Functional similarity is strong. Reuse still needs the evidence listed below.
Why they matched
| Component | Contribution to score |
|---|---|
| System outputs | +0.520 |
| Same AI classification | +0.037 |
| Total | 0.557 |
Most distinctive shared terms: searchable index, index, searchable, entities
| Field | Plum Island handwriting and text extraction from scanned pdfs | Named Entity Recognition |
|---|---|---|
| Agency |  USDADepartment of Agriculture |  DOJDepartment of Justice |
| Bureau | Marketing and Regulatory Programs | Department of Justice / FBI |
| Stage | Pre-deployment | Deployed |
| Topic | Science | Information Technology |
| AI classification | Natural language processing | Natural language processing |
| Sourcing | Not reported | Developed in-house |
| Vendor (reported) | Not reported | Not reported |
| System name | Not reported | Redacted for cybersecurity purposes. |
| Problem | To reduce manual intervention and automate searchability of Plum Island lab manuals and scientific documents. | Entity Extraction |
| Outputs | Searchable index of key phrases, entities, authors, and additional fields relevant to the documentation. | Resolved entities in a searchable index. |
| Benefits | Time savings that could translate to more personnel cycles on scientific needs as opposed to manual labor searches. | Faster FBI operations |
| Data description | Not reported | The case owner relied on DOJ AI governance practices to select and prepare data, as well as evaluate performance. |
| PII | Not reported | Yes |
| ATO | Not reported | Yes |
| Public code | Not reported | Not reported |
Exploratory reuse assessment
What the inventory can and cannot show| Functional similarity | strong (0.56) | Some support |
| Inputs and outputs | Both describe outputs; compare below | Some support |
| Mission or administrative function | Different topic areas | Weak or missing |
| Documentation available | Code: none linked · System names: Not in inventory | Weak or missing |
| Technical uncertainties | AI classification: same; sourcing: unknown / in_house | Verify outside inventory |
| Data-access constraints | Data descriptions: Not in inventory / Reported | Verify outside inventory |
| Privacy and security | PII: blank / yes · ATO: blank / yes | Verify outside inventory |
| Legal or policy constraints | Not in inventory. Requires independent verification. | Verify outside inventory |
No pair is marked as a validated reuse opportunity. That needs technical, legal and operational evidence beyond this inventory.
Questions for a human reviewer
- Do USDA and DOJ solve the same problem for the same kind of user, or only use similar words?
- Are the inputs (data types, formats, volumes) compatible?
- At least one system reports PII. Would sharing require a new privacy assessment or data-sharing agreement?
- Does either system's Authorization to Operate cover use by another agency?