What is “black-box AI”

A simple explanation of black-box AI and why it matters.

Black-box AI is when you can see the input and the output, but not how one became the other. Not because it’s deliberately hidden, but because the process is too complex to explain in a useful way.

For example, you apply for a job and get rejected before a human ever sees your CV. The system has learned patterns from past hires, but there’s no clear reason you can point to.

That’s where it starts to matter. When AI is used in hiring, credit decisions, or healthcare, the issue isn’t just the outcome, it’s not being able to question it.

Most of the time, people don’t think about this until something feels off. And when it does, the question is simple: do we trust decisions we can’t explain?