HHS · no agency ID · record R0981

Detection of AI-Generated Reviewer Critiques

Department of Health and Human Services · HHS/NIH

Overview

Development stage: PilotSource: b) Pilot – The use case has been deployed in a limited test or pilot capacity.
High-impact designation: Not high-impactSource: c) Not high-impact
Impact justification: Not reported
Start date: 2020-01 (month precision)Source: Jan-20
Withheld from public reporting: NoSource: a) No

Mission

Topic area: Science
Operational functions: No function tag matched Derived by keyword rules; see methodology

Problem the AI is intended to solve

This tool detects if a given written critique was generated by AI tools, such as large language models (LLMs), as the reviewers are not allowed to write their critiques using these AI assistants. Flagged critiques are examined by NIH staff and discussed with the reviewers.

Expected benefits

NIH policy mandates that reviewers produce independent analyses of grant applications based on their expertise and knowledge. The use of genAI to produce a written critique of a grant application is in violation of NIH policy and fails to provide an independent assessment of the application.

System outputs

This tool outputs classifications of whether the reviewer critique was likely produced by Generative AI or not.

Technology

AI classification: Natural language processingSource: Natural Language Processing: AI that processes, interprets, and shares information in human language.
System name(s): CSR Detection of AI Gen Reviewer Critiques (CSRDAIRC)
Custom-developed code: Yes
Public source code: Not reported

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: Yes
Demographic features: k) None of the above
Training and evaluation data: critiques written by the reviewers, AI-generated critiques, Amazon book reviews, and AI-generated book reviews
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.

Pre-deployment testingNo answer reported

AI impact assessmentNo answer reported

Independent reviewNo answer reported

Ongoing monitoringNo answer reported

Operator trainingNo answer reported

Fail-safeNo answer reported

Appeal processNo answer reported

User and public consultationNo answer reported

Potential impacts and how they were identified

Not reported