https://trendland.com/illustrated-magazine-covers/
This drawing robot will not steal your job. It draws layers on tracing paper, each layer with a different ink darkness to correspond to the area with corresponding color intensity / brightness on the magazine cover. Then the tracing paper are stacked together to show the recreated cover.
Potential for machine learning: I don’t know if there’s database for this, but if there is, there should be tools online available to train an model that tells if an image is a magazine cover or not, so when the viewer submits/ downloads an image and feeds it to the program, the single purpose robot could decide whether to draw it or not.
Technical details: I don’t know how to go about building a drawing robot, but google says they could look like this. Splitting the magazine cover into layers should be doable by treating image as array of pixels, averaging them to make it lower resolution, and determine which layer each square should belong to.
For the look, I’m thinking chinoiserie if its using a traditional Asian calligraphy brush; the robot can also draw using a stamping motion with stamp / sponge, and stamp out more defined squares.
The drawing process may end up slow, so probably viewer would simply queue the magazine cover they want to have drawn, or help put the drawing together.