Contribution: Industrial Research Assistance Program – Contributions to Firms
Armilla AI Inc.
$75,000.00
Developing an assessment framework to evaluate bias of Generative AI in Human Resources (HR) Systems
FY 2022-23 · 2023-02-13
Summary
Generative AI can be used in HR systems to perform a myriad of functions: it can rank candidates, create job descriptions, generate ideal skills for a job description, create interview questions, as well as many other HR tasks. However, generative AI systems have been shown to contain implicit and explicit biases, hate, and other negative traits based on the data they have been trained upon. Testing these systems for bias is difficult but will become a requirement, starting with the New York City Local Law 144 and the upcoming EU AI Act. In this proposal, we aim to develop a test framework for testing these systems and making them more robust.
Department
Program
Industrial Research Assistance Program – Contributions to FirmsIndustrial Research Assistance Program-Contributions to Firms (linked with high confidence)
Recipient
Location
Toronto|Toronto, Ontario, Canada
Recipient Type
For-profit organization
Business Number
714072873
Type
Contribution
Source
Proactive Disclosure - Grants and Contributions
As published
Agreement End Date
2023-05-31
Agreement Number
1003397
Agreement Start Date
2023-02-13
Agreement Title
Developing an assessment framework to evaluate bias of Generative AI in Human Resources (HR) Systems
Agreement Title (French)
Développement d’un cadre d’évaluation des biais de l’IA générative dans les systèmes de ressources humaines (RH)
Agreement Type
Contribution
Agreement Value
$75,000.00
Amendment Number
0
Coverage
IRAP Contributions to Firms projects stimulate wealth creation for Canada through technological innovation. | Les projets des Contributions aux entreprises stimulent la création de richesse au Canada grâce à l’innovation technologique.
Description
Generative AI can be used in HR systems to perform a myriad of functions: it can rank candidates, create job descriptions, generate ideal skills for a job description, create interview questions, as well as many other HR tasks. However, generative AI systems have been shown to contain implicit and explicit biases, hate, and other negative traits based on the data they have been trained upon. Testing these systems for bias is difficult but will become a requirement, starting with the New York City Local Law 144 and the upcoming EU AI Act. In this proposal, we aim to develop a test framework for testing these systems and making them more robust.
Description (French)
L’IA générative peut être utilisée dans les systèmes RH pour remplir une myriade de fonctions : elle peut classer les candidats, créer des descriptions de poste, générer des compétences idéales pour une description de poste, créer des questions d’entretien ainsi que de nombreuses autres tâches RH. Cependant, il a été démontré que les systèmes d’IA générative contiennent des préjugés implicites et explicites, de la haine et d’autres caractéristiques négatives basées sur les données sur lesquelles ils ont été formés. Il est difficile de tester la partialité de ces systèmes, mais cela va devenir une obligation, à commencer par la loi locale 144 de la ville de New York et la future loi européenne sur l’IA. Dans cette proposition, nous visons à développer un cadre d’essai pour tester ces systèmes et les rendre plus robustes.
Expected Results
The purpose of IRAP Contributions to Firms is to support research, development, adoption and/or adaptation of innovative or technology-driven new or improved products, services or processes in Canada up to their commercialization
Expected Results (French)
Le PARI, par le biais des Contributions aux entreprises, vise à appuyer la recherche, le développement, l’adaptation et/ou l’adoption au Canada de produits, de services ou de processus novateurs ou axés sur des technologies nouvelles ou améliorées, et ce jusqu’à leur commercialisation
Federal Riding
University--Rosedale
Federal Riding (French)
University--Rosedale
Federal Riding Number
35110
NAICS Code
511211
Organization Code
nrc-cnrc
Organization
National Research Council Canada | Conseil national de recherches Canada
Program Name
Industrial Research Assistance Program – Contributions to Firms
Program Name (French)
Programme d’aide à la recherche industrielle – Contributions aux entreprises
Program Purpose
The purpose of IRAP Contributions to Firms is to support research, development, adoption and/or adaptation of innovative or technology-driven new or improved products, services or processes in Canada up to their commercialization
Program Purpose (French)
Le PARI, par le biais des Contributions aux entreprises, vise à appuyer la recherche, le développement, l’adaptation et/ou l’adoption au Canada de produits, de services ou de processus novateurs ou axés sur des technologies nouvelles ou améliorées, et ce jusqu’à leur commercialisation
Recipient Business Number
714072873
Recipient City
Toronto|Toronto
Recipient Country
CA
Recipient Legal Name
Armilla AI Inc.
Recipient Postal Code
M6G 1S3
Recipient Province or Territory
ON
Recipient Type
F
Reference Number
172-2022-2023-Q4-1003397
