How Does 3 types of bias in AI | Machine learning Work?

Imagine you're sorting toys into two boxes, one for big toys and one for small toys. But if most of the toys you see are big red cars, your box might end up with more big red cars than it should. That’s what happens in AI when there's a type of bias called selection bias.

Sometimes, the person who picks out the toys, like your friend, only brings you red toys, even though they come in all colors. This is similar to another kind of bias in AI called confirmation bias. The AI thinks red is the best color because that's what it keeps seeing.

Now imagine the toy box itself has a favorite, it always moves big toys to the front and pushes small ones behind. That’s like algorithmic bias, where the way the AI works makes certain results more likely, even if all the toys are fair.

These three types of bias help explain why sometimes an AI might choose red cars over blue trucks or think most toys are big, not because they're wrong, but because that's what they've mostly seen! Imagine you're sorting toys into two boxes, one for big toys and one for small toys. But if most of the toys you see are big red cars, your box might end up with more big red cars than it should. That’s what happens in AI when there's a type of bias called selection bias.

Sometimes, the person who picks out the toys, like your friend, only brings you red toys, even though they come in all colors. This is similar to another kind of bias in AI called confirmation bias. The AI thinks red is the best color because that's what it keeps seeing.

Now imagine the toy box itself has a favorite, it always moves big toys to the front and pushes small ones behind. That’s like algorithmic bias, where the way the AI works makes certain results more likely, even if all the toys are fair.

These three types of bias help explain why sometimes an AI might choose red cars over blue trucks or think most toys are big, not because they're wrong, but because that's what they've mostly seen!