CDC Vault and Stacks Metadata Extraction

Overview

Development stage

DeployedSource: c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.

High-impact designation

Not high-impactSource: c) Not high-impact

Impact justification

Not reported

Start date

2025-04 (month precision)Source: Apr-25

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Administrative Functions

Operational functions

No function tag matched Derived by keyword rules; see methodology

Problem the AI is intended to solve

We are attempting to speed up the process of generating a digital metadata record for objects that will be curated and stored into either CDC Public Access Platform (Stacks) or CDC Vault. These two systems are built using the same software stack but one is for public data and the other is for non-public data. To create a metadata record solely with a human, the process takes about an hour per document. We are looking to improve the process to use AI to prepare the metadata record and reduce the human time to under 5 minutes. A secondary objective is to have a non-human process for the non-public data that will go into CDC Vault.

Expected benefits

There are two primary paths and uses for the AI assisted pre-processing. The first is to improve the speed and effectiveness of human catalogers/librarians. Long term, we need to be able to process more data and require AI to improve this process so that humans are only working on critical steps and validation of the AI. This process is going from 60 minutes per document to <5 minutes per document. The second is to process federal records prior to a record being entered into CDC Vault and copied to NARA. This process will not have a human review as the final disposition is not public but we need to process a large number of files (100s of thousands to millions). This is simply not realistic to do via humans so this is a novel opportunity.

System outputs

The AI will return up to 41 metadata elements (eg Title, Author, Subject, Description, Funding Source, Geographical Local).

Technology

AI classification

Agentic AISource: Agentic AI: AI systems that perform tasks or make decisions autonomously with minimal human intervention.

System name(s)

EDAV (Enterprise Data Analytics and Visualization Platform)

Custom-developed code

No

Public source code

Not reported

Sourcing

How it was built

Contract and in-houseSource: c) Developed with both contracting and in-house resources

Vendor (as reported)

Not reported

Vendors (standardized)

None on the standard list

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII

No

Privacy Impact Assessment

Not reported

Authorization to Operate

Yes

Demographic features

k) None of the above

Training and evaluation data

Not reported

Federal Data Catalog

stacks.cdc.gov ↗

Governance

0 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Potential impacts and how they were identified

Not reported