The Nov. 10 Business article “Applying for a job? A face-scanning algorithm could be the gatekeeper.” described HireVue’s claims that its artificial-intelligence evaluation of video interviews operates with “superhuman precision and impartiality.” Using AI, however, will not eliminate bias in hiring.

An AI algorithm bases its judgments on similarities with the data it’s given. For example, an AI algorithm used to evaluate arrestees considered black people more likely to reoffend than white people. This was because the data used to train the AI had a much higher rate of recidivism for blacks, possibly because of systematic bias. Instead of eliminating prejudices, AI can obfuscate them, masking them behind a veneer of mathematical certainty.

HireVue claims it can eliminate such “adverse impacts” by drawing on videos of diverse people. However, the problem cannot be solved this easily. Personality traits that lead to promotion, such as extroversion, will be selected for, even if they are not beneficial to an employee’s performance on the job.

We must stop thinking of AI as a panacea for human subjectivity. Artificial intelligence will do only what we tell it to do.

Daniel Fu, Great Falls

As someone who has spent many years hiring people, the Nov. 10 Business article about using artificial intelligence in hiring (as a candidate said) “made my skin crawl.” Amazon and Netflix now use algorithms to determine something much simpler than who is right for the job. They show me what they think my next purchase or movie rental would be, and they are dead wrong more than 95 percent of the time. More than 95 percent of my hires have been terrific employees; I doubt artificial intelligence could do better.

Roslyn Lang, Bethesda