It is less a 3D feature rather then a render feature (vector to bitmap render). Your example (a 2D vector shape) is a good demonstration.
The simplified render technique assumes the middle of a pixel determines the overall color of it. This is a valid assumption within a shape but at the edges it leads to these “stairs” artifacts. These artifacts are noticeable on low resolution screens and “thin” lines.
To make this appear a bit more natural to our eyes a pixel can be covered by multiple different shapes. The resulting color is a mix of these shapes. In the above example a mix of black and white shapes. Dependent on the ratio the pixels gets different shades of gray.
There are multiple different methods to do achieve AA. All of them increase the render time.
Alternatively you can increase the resolution. But this depends on hardware. Higher resolution (with same physical size) usually means more render time too.