Where AI bias comes from

AI bias isn’t a glitch bolted on at the end. It’s a reflection of the data the model learned from. Because models are trained on human writing, they absorb human patterns, including the unfair ones. Knowing where the bias comes from makes it easier to spot.

Where it shows up

Ask for “a nurse” or “a CEO” and a model may lean on the stereotypes that were common in its training text. Quality is uneven too: models tend to perform better in widely represented languages and contexts than in under-represented ones. And a model can present a partial, culturally narrow view with complete confidence, as if it were the whole picture.

What you can do

You can’t retrain the model, but you can change how you use it. Be specific, so the defaults don’t do the talking. Ask for the perspectives you suspect are missing. Treat the output as a draft to review; bias is one more reason the human stays the editor. And in any decision that affects a person’s prospects, a hiring shortlist being the obvious example, never let a model be the unaccountable decider.

A model’s fluency can make a biased answer sound authoritative. Sounding sure is not the same as being fair.

People deserve to know when AI shaped something that affects them. That is the disclosure question, covered in when to tell people you used AI.