Pregnancy discrimination in the AI era

On Behalf of | Jan 24, 2024 | Pregnancy Discrimination |

A majority of employers in California and around the country use automated tools like artificial intelligence algorithms when they screen job applicants, which some experts believe will lead to a sharp rise in pregnancy discrimination. AI algorithms narrow down a field of candidates by weeding out unsuitable applicants, but the criteria they use could also exclude pregnant applicants or applicants who recently had children. To prevent this kind of employment discrimination, the Equal Employment Opportunity Commission launched the Artificial Intelligence and Algorithmic Fairness Initiative in 2021.

Absenteeism and sick days

The AI algorithms used by American employers often identify job candidates who are frequently absent or take a lot of sick days as undesirable. This could lead to pregnancy discrimination because pregnant women have frequent doctors’ appointments and are sometimes unable to work. To help organizations to avoid these missteps and comply with federal employment laws, the EEOC offers guidance on algorithmic impartiality and technical assistance.

Proxy discrimination

However, following EEOC guidelines may not be enough to eliminate unintended discrimination in the hiring process. When AI algorithms are adjusted to not eliminate candidates because of absenteeism or frequent sick days, they use other criteria like gaps in the applicant’s employment history. When this also leads to applicants being treated unfairly, it is called proxy discrimination. U.S. privacy laws do not currently protect data dealing with reproductive health, which is why experts think proxy discrimination against pregnant job candidates will become a thorny problem in the years ahead.

Unfairness in the hiring process

Federal law requires prohibits most companies from discriminating against pregnant job applicants and workers, but the increasing reliance on AI algorithms during the hiring process may be causing unintended discrimination. These algorithms often identify undesirable candidates by sorting them according to the number of sick days they take or how often they do not show up for work, but these criteria may also eliminate pregnant candidates and new mothers. Changing the criteria may not solve the problem because of a phenomenon called proxy discrimination, so the EEOC has launched an initiative to keep hiring processes fair in the age of AI.