NLP Automated Referral

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

Presumed high-impact, determined notSource: b) Presumed high-impact, but determined not high impact

Impact justification

First, the NLP Automated Referral tool produces non-binding recommendations regarding which NIGMS program officer is most appropriate to manage an incoming grant application, along with two alternative suggestions. Program officers must actively accept the assignment or refer the application to a more appropriate program officer, and they have full discretion to ignore or override the AI tool’s recommendations. The system’s output is therefore not a principal basis for any legal or binding decision; it is one of several informational inputs to an internal workflow choice made by human staff. Second, all funding and programmatic decisions are made through established peer review and programmatic processes, governed by existing NIH/NIGMS policies and human judgment. Final funding decisions are made by the NIGMS Director in consultation with the NIGMS Advisory Council, the NIGMS Deputy Director, and the NIGMS Division Directors, not the individual program officers who manage the applications. As a result, the AI’s output does not directly affect an individual’s or organization’s access to Federal funding or other critical government resources or services, nor does it alter anyone’s legal status or rights. Finally, the NLP Automated Referral tool does not fall into any of the categories of AI use cases identified in Section 6 of M-25-21 that are automatically designated high-impact. It is a routing tool used for internal staff portfolio management.

Start date

2020-08 (month precision)Source: Aug-20

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

Referring applications manually is tedious and time consuming.  Using an automated approach allows staff to focus their time on more difficult tasks.

Expected benefits

Automated referral allows NIGMS to retain institutional referral knowledge by training on historical data, eliminates delays in referral by assigning applications as soon as they come in, and reduces burden on staff members and allows them to allocate more of their time to other high value tasks.

System outputs

Input: IMPAC II application data, including titles, abstracts, narratives and specific aims. Output: Top three most relevant ICs and POs.

Technology

AI classification

Classical / predictive MLSource: Classical/Predictive Machine Learning: Models trained on data to make predictions or classifications based on identified patterns or relatio…

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 come from the internal 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