NIGMS Azure Open AI

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

2024-09 (month precision)Source: Sep-24

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

There are a number of situations in which administrative activities can be augmented by generative AI, especially when classification of documents is needed but no training data exist.

Expected benefits

We are using large language models (LLMs), in particular Open AI's models, for business process improvement.  We have used these models for visualization of a grant portfolios as well as numerous classification problems: IC prediction, clinical trial prediction, research area prediction, etc.  These models allow us to classify documents through simple prompt engineering rather than the laborious process of creating a custom training set from scratch.  These models also allow us to reduce the number of applications that humans need to review from tens of thousands of applications to mere hundreds or fewer for a number of tasks.

System outputs

Input: Text from various components of NIH grant applications. Output: Open AI chat completions (text) or text embeddings (vectors of numbers).

Technology

AI classification

Generative AISource: Generative AI: AI that generates new or synthetic content (e.g., images, videos, audio, text, code).

System name(s)

Not reported

Custom-developed code

Yes

Public source code

It is not publicly available.

Sourcing

How it was built

Developed in-houseSource: b) Developed in-house

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

No

Demographic features

k) None of the above

Training and evaluation data

All data used for this project is internal to NIH, mostly administrative data from the NIH IMPAC II database.

Federal Data Catalog

Not reported

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