Georgia Tech launched a new course, Art and Generative AI, to smash AI myths and boost real creativity. The class challenges the idea that AI is humanlike, stressing that machines just mimic patterns without true understanding or care.
The course dives into AI’s nuts and bolts — from perceptrons to transformer models like those behind ChatGPT — while blending in hands-on art practice. Students mess with AI limits by using small datasets and scaled-down models to provoke glitches and hallucinations. This trains them to become conscious co-creators, taking back control from AI.
Starting spring 2025, the class team-taught with artist Mark Leibert, teaching everything from charcoal drawing to oil painting alongside AI ethics. Students crafted AI-based art with respect for originality and authorship. 
They also hooked up EEG headsets to turn brainwaves into music and images, inspiring live improvised dance performances reacting to AI-generated sound.
“We reject the illusion that machines fully master everything and put student expression first. In doing so, we value uncertainty, mistakes and imperfection as essential to the creative process.”
The course tackles AI’s environmental costs, teaching students to cut wasteful chatbot prompts and lower carbon footprints.
“The artwork isn’t generated by AI, but it’s reimagined by students.”
By 2025-26, the program will extend to community collaborations in Atlanta’s art scene, with local artists co-teaching to merge artistic practice and AI.
This course aims to prep students to take on real-world engineering challenges. It drives home that ethical AI needs good data and human authorship to avoid bias and errors.
