AI facial recognition is often marketed as precise and objective but when it gets things wrong, the consequences can be serious, even life-altering.
At its core, facial recognition works by comparing key facial features distance between the eyes, shape of the jaw, contours of the nose to massive image databases. The problem is that similarity is not identity, and humans are far more alike than these systems are willing to admit.
One of the greatest dangers is false positives when the system confidently identifies the wrong person.
There have already been documented cases where people were:
- Wrongfully arrested
- Detained for hours or days
- Investigated for crimes they did not commit
—all because an algorithm matched their face to a grainy image or low-quality surveillance photo.
Artificial intelligence got a man wrongly arrested at the Peppermill casino. He’s now suing a Reno police officer in federal court. Watch the video below:
