Robot drawings provide a portrait of AI’s future

Visual artist Sougwen Chung collaborates with robots. It began several years ago when she used open source software to create a robotic arm that would copy her every brush stroke, thus painting alongside her. Named Drawing Operations Unit: Generation 1 or DOUG_1 for short, it soon emerged that the arm had mechanical and technological limits that rendered its imitation imperfect.

This imperfection, says Chung, is what made the collaboration more interesting: “The machine was interpreting my lines but not perfectly, and I was forced to adapt in real time. It’s far from the omnipotent AI that we’ve been told to believe in.”

Since then there have been two more DOUGs. The second was a neural network that learned how to create art like Chung’s after processing more than 20 years of her finished paintings, sketches and other works.

But Chung wanted to push the envelope and the third iteration was not one robot but a cluster of 20 that worked as a collective, not only with the artist but with all of New York city. These robots were able to “see” the city by tapping into publicly available webcams of people walking, cycling and driving. With this as inspiration, Chung and her robot collective are redefining landscape painting. Watch her TED talk to learn more about this fascinating blend of art and AI.

[ https://www.ted.com/talks/sougwen_chung_why_i_draw_with_robots/ ]