AEDT BIAS AUDIT
Prepared by Conductor AI on 04/04/2024 in accordance with NYC Local Law 144
AEDT Covered Fit Score
Opening up the future
We empower teams to achieve their goals. By building software that makes working together easier, we can help every team achieve their most ambitious goals. It’s our chance to accelerate human progress. We’re off to a great start—fast-growing companies like Uber and Pinterest already run on Phenom. We have a considerable mission and an even bigger opportunity: to change the way teams work together. We’d love your help.
Summary of Findings:
The audit results shown below were produced on 04/04/2024, for, regarding their use of the following tools: Fit Score.
This bias audit does not represent Conductor AI’s or ’s opinion that Fit Score is an automated employment decision tool (AEDT). The purpose of the audit is to analyze Fit Score as if it were an AEDT to enable compliance with NYC Local Law 144.
The audit below was performed using historical data from between 01/01/2023 and 12/31/2023 gathered from multiple employers provided by to the independent auditor, ConductorAI.
The dataset provided by included jobs distributed across the United States, multiple employers, multiple job families, and distinct talent pools for each job. As a result, the initial scores have been normalized across all job types to produce final scores that are comparable across the entire dataset for the purposes of this audit. This normalization process created a final score for a candidate by multiplying their initial score by the median score of the dataset divided by the median score of their job type.
It is the independent auditor’s opinion that the provided data is representative, and “Scoring Rate” is most appropriate for assessment. Scoring Rate is the percentage of candidates within a given group whose score is above the median score of the entire dataset. This median score across all candidates is 1.5863. After normalization, the median score across all candidates was .5. Impact Ratio is calculated by dividing the Scoring Rate of a given group by the highest Scoring Rate within the category.
The dataset given to ConductorAI consisted of 3,064,538 candidates. 232,786 candidates had sex data but unknown race data. 339,455 candidates had race data but unknown sex data.
This bias audit does not represent Conductor AI’s or ’s opinion that Fit Score is an automated employment decision tool (AEDT). The purpose of the audit is to analyze Fit Score as if it were an AEDT to enable compliance with NYC Local Law 144.
The audit below was performed using historical data from between 01/01/2023 and 12/31/2023 gathered from multiple employers provided by to the independent auditor, ConductorAI.
The dataset provided by included jobs distributed across the United States, multiple employers, multiple job families, and distinct talent pools for each job. As a result, the initial scores have been normalized across all job types to produce final scores that are comparable across the entire dataset for the purposes of this audit. This normalization process created a final score for a candidate by multiplying their initial score by the median score of the dataset divided by the median score of their job type.
It is the independent auditor’s opinion that the provided data is representative, and “Scoring Rate” is most appropriate for assessment. Scoring Rate is the percentage of candidates within a given group whose score is above the median score of the entire dataset. This median score across all candidates is 1.5863. After normalization, the median score across all candidates was .5. Impact Ratio is calculated by dividing the Scoring Rate of a given group by the highest Scoring Rate within the category.
The dataset given to ConductorAI consisted of 3,064,538 candidates. 232,786 candidates had sex data but unknown race data. 339,455 candidates had race data but unknown sex data.
Widget title goes here.
Card title goes here.
A brief description highlighting the key message or action users can take.