New York City bias audit

Background

In response to the enactment of Local Law 144 by the New York City Council in 2021, governing “automated employment decision tools,” and subsequent rules adopted by the New York City Department of Consumer and Work Protection (DCWP) in April 2023, enforcement officially commenced on July 5, 2023. This legislative framework introduces responsible AI provisions emphasizing accountability, fairness, transparency, and explainability. Testlify provides the information on this webpage to foster public understanding of our Matching Model, irrespective of its coverage under New York City Local Law 144 of 2021.

Testlify assessment model

Testlify, an innovative talent assessment tool, is uniquely designed to serve companies and HR professionals by facilitating assessment invitations for candidates on its platform.

Employing advanced machine learning techniques within its cloud-based platform, Testlify ensures an unbiased evaluation process. The resulting match score, presented as a percentage out of 100, offers decision-makers valuable insights into the compatibility between a candidate and a job position. This dynamic score is finely tuned based on the specific pairing of candidate and job position, as well as updates to candidate or job information.

Beyond overall assessment scores, Testlify provides a granular view of candidate performance for individual skills and across all tests. This detailed analysis assists companies in gauging a candidate’s relevance to the specific positions they are hiring for. By showcasing both overall and skill-specific scores, Testlify empowers HR professionals with comprehensive insights for informed decision-making.

Testlify provides valuable data insights to support Human Resource (HR) processes, emphasizing that the model does not replace human decision-making. Regardless of match strength, candidates can apply for any open position, and HR recruiting users retain the autonomy to hire as they see fit.

Inputs for the model

The Testlify assessment model analyzes data derived from tests curated by recruiters and the responses submitted by candidates. Testlify assigns scores to candidates based on their performance in these tests. Employers have the flexibility to create tests tailored to the specific skills relevant to the role they are hiring for. These tests may encompass multiple questions designed to evaluate various skills.

These inputs come from individuals, such as candidates or recruiters, or databases/systems like an employer’s applicant tracking system (ATS) or human resources information system (HRIS). Input data may be accessed from cloud-hosted systems or integrated with Testlify’s platform, allowing employers to provide up-to-date information. Testlify’s Privacy Policy provides additional details on personal data processing.

Viewing match score outputs

Utilizing the responses provided by candidates, the model generates an assessment score for each candidate. Testlify presents this score to employers, accompanied by a detailed analysis for each test, offering profound insights. The platform not only showcases the overall match score but also provides clear explanations of how the model computed the score, including a concise breakdown

Reviewers can use the model output to understand the specific skills, titles, work, and job title relevance of each candidate for a position. They can visit the candidate detail page for breakdowns and check the performance. The display allows flexibility in viewing candidates and taking actions regardless of the model output.

Bias audit

Testlify has undergone a third-party bias audit of the Matching Model, conducted by the independent auditor. The most recent audit, completed on June 16, 2023, is available for review. This audit aligns with the requirements of New York City Local Law 144 of 2021.

Data retention

Testlify’s data retention policy for the personal data of candidates, including assessment Model data, is detailed in our Privacy Policy. Refer to the “Data Retention” section of the policy. Employees seeking information on internal data retention policies may contact [email protected]