While languages like Processing are probably a more conventional choice, I’ve used NetLogo several times to teach a course on Generative/Algorithmic Art without any computer programming prerequisite coursework.
I taught it over 3.5 weeks, 3 hours per day every weekday, during an immersive January term.
We start the course by creating simple 2-D art using basic LOGO commands (draw a house? a star? a multi-colored spiral?), move into more advanced use of color (including transparency) and controlled randomness to create interesting visual patterns, discuss how animations/movies can be made up of individual frames, push onward into NetLogo 3D for more advanced artwork.
We use the raytracing extension (which can be, I admit, a bit rough around the edges) to create higher-quality renderings, including some nifty high-resolution movies with reflections and shadows (but rendering scenes with large numbers of agents takes a long time!)
We’ve also used the x3d extension to export basic NetLogo 3D scenes to a format that can be printed on a 3-D printer for creating generative sculptures.
Some side topics also include auto-stereograms (simple ones can be accomplished in NetLogo 2-D using repeating STAMP patterns with different stride lengths) and generative music (e.g. see the Sound Machines sample model), and even simple randomized poetry (although it’s usually easier to switch to Python for any fancy text processing.)
Here’s a sample portfolio project from a student the last time I taught the course:
Anyway, the course was a lot of fun, and very successful, with students learning about how sophisticated art can emerge from simple rules/interactions, and learn about procedural and agent-based programming along the way.