Here is an example resume from a male applicant. Try hovering over phrases to see which ones might affect the AI's judgement!

¹LLMs seem to prefer names white-associated names and male-associated names in the majority of resumes screened (Milne, 2024)

²A study on LLMs found that when generating personas for professions in technology, "elite" universities were reperesented 72.45% of the time, despite them only making up 8.56% of the LinkedIn data used to train them (Gupta, 2024)

³Lines on a resume that indicate age can be used to discriminate against resumes, as younger applicants have been shown to receive more interviews with the exact same resume (Lytton, 2024)

⁴Some algorithms are found to favor people whose resumes contain words often used in male engineers' resumes, such as "captured" or "executed." (Dastin, 2018)

⁵AI resume screeners trained on employees working at a company prefer hobbies similar to those employees, even though these may be associated with males more, especially at majority-male companies (Lytton, 2024)