AI Artists Logo.png

Pindar Van Arman

Pindar Van Arman invented CloudPainter, and has created robots that paint for more than ten years using AI, robotic arms and 3D printing.

Pindar Van Arman - Cloud Painter.jpg


Pindar Van Arman is an award-winning AI Artist who builds, codes and experiments with robots that paint in a variety of styles. Pindar was a winner of the Robot Art Competition for his home-made robots that use AI to paint artworks.

Van Arman’s exploration into robot art began by following his curiosity, attempting to answer the question: where does creativity come from? To try to understand this universal question, he dissects his own artistic process and teaches robots to paint.

Van Arman’s impressive and award-winning works include CloudPainter, an autonomous robot that anyone with an Internet connection can control to produce artworks. He has been featured in many interviews, TV episodes, and conferences.

Van Arman’s first machine drew simple lines with a paint brush. Now his most recent robot uses a custom 3D printed paint head, two robotics arms, machine learning, and artificial intelligence to create its own original compositions and artwork, similar to the way human artists operate.

Not just simple assistants, his machines create genuinely original works that have impressed even NYC Art Critic Jerry Saltz, who said his robots “have good taste” - and look like they could be made by the human hand. You can see a humorous roundup of Saltz’ critiques of several machine-made paintings, including Van Arman’s in the video below.

Van Arman’s TED talk:

Watch Van Arman explain his process of teaching creativity to robots in his TED talk below.

Selected works:

Pindar Van Arman - Painting.jpg
Pindar Van Arman - Painting.jpeg
Pindar Van Arman - Ghostfaces Shadow.jpg

Studio shots:

Pindar Van Arman - Painting in Progress.jpeg
Pindar Van Arman - In the studio.jpg

Van Arman on using artificial intelligence:

How has AI impacted your creative practice?

“Attempting to teach my machines how to paint has forced me to reverse engineer my own creative process, and as a result come to a more profound understanding of it,” Van Arman explains.

“With each painting my robots create, I am attempting to teach them how to independently make all the aesthetic decisions that I make when I paint. This has been a fascinating process because in order to teach my own creativity to a machine, I am forced to deconstruct the process into logical steps. Sometimes I am successful in my attempts, and other times I fail. But when I do find an algorithm that performs a specific artistic function, and it performs that function well, I can not help but wonder if that is what is happening in my own mind. This process of deconstructing creativity gives me insight into how my own creativity works – and a clearer understanding of how my own mind works. What is cooler than developing a deeper understanding of ourselves by becoming more familiar with our creative process?”

What excites you most about AI as an artist?

“Can AI make art? Of course not. Not yet, anyways,” Pindar explains. “But another interesting question is this: can AI be creative? The answer is yes, it absolutely can. If you doubt it, you need look no further than the recent surge of AI artists around the world that are demonstrating many of the creative things that AI can do. As an artist working with AI for more than fifteen years, today is a really exciting time to be a part of that community.”

What specific AI / machine learning technologies does you use?

“One of my recent works used at least twenty-six distinct algorithms, some borrowed, others stolen, and some even created,” Van Arman explains. “I love finding ways to utilize k-means clustering to organize things such as colors. I also like implementing the various algorithms I find in OpenCV such as Hough Lines to help find edges, and Haar Cascades to recognize and help paint faces. Recently, things have exploded creatively with deep learning. You will find style transfer (CNNs) in many of my paintings and I utilize Generative Adversarial Networks (GANs) whenever I want to create imagery from nothing. The most important algorithm for me, however, is the feedback loop. The fact that my robots are watching every stroke, taking a step back and analyzing how the stroke impacts the artwork, then using that analysis to help form the next stroke is – which is very similar to my own creative process. In fact, I think it is similar to most painter's creative process. We make a mark, take a step back, look, then make another mark. We do this over and over again in a continuous feedback loop until a painting is complete, and so do my robots.”

Awards, Exhibitions and Talks:

  • NeurIPS AI Art Gallery 2018 - Curated Gallery featuring Machine Learning for Creativity and Design

  • out of office. when A.I. and robots work for us - Museum der Arbeit, Hamburg, Germany - Nov 2018 - Mar 2019

  • Robotic Artist in Residence at Cook Children's Medical Center - Fort Worth, Texas

  • DC DataCon 2018 - Artificial Creativity - Washington D.C.

  • Tysons: Then and Now - Solo Exhibition - Sep-Nov 2018 - Tysons, VA

  • RobotArt 2018 - First Place Prize in third year of the $100,000 International Competition.

  • ART.IFICIAL - Website and Book by GumGum exploring artificial creativity and its role in art.

  • Aspen AI Conference - The First Sparks of Artificial Creativity - Berlin, Germany - 2018

  • MBN Y Forum 2018 - Seoul, Korea - 2018

  • DC DataCon 2017 - The Threshold Between Human and Computational Creativity - Washington DC. - 2017

  • IROS 2017 - Towards and Artist in the Lab - Invited Speaker - Vancouver, BC- 2017

  • RobotArt 2017 - Top Technical Contribution and Third Place Prize in second year of competition - 2017

  • NVIDIA GTC 2017 - Featured in Keynote - 2017 

  • CISR 2017 - Speaker - Oak Ridge National Labs, 2017

  • Elastic Cause Award - Shortlist - Demo at Elastic{ON} - 2017

  • Microsoft Surface Innov8r - Short Feature - 2017

  • TEDx Talk - Teaching Creativity to Robots - 2016

  • RobotArt 2016 - Second Place Prize in Inaugural Year of the International RobotArt Competition - 2016

  • - Live ongoing interactive crowd painting performance piece and experiment - 2013-Present

  • Google Made w/ Code Launch Party - NY - 2014

  • Barbican's Dev Art Competition - Shortlisted - 2014

  • Form Follows Software - Group Show - Arizona State University - 2012

  • Gallery One - The Corcoran - Student Group Exhibition - Wash DC - 2010

  • Kids Collaborate with Robot Artists - Exhibition with Max Chandler - The Cartoon Network Headquarters, LA - 2010

  • Beyond Warhol in the 21st Century: A Post-Pop MetaRomantic Retrofit - BAG Gallery - Group Show - Brooklyn - 2010

  • FLUX 09 - Group Show - Wash DC - 2009

  • The 12"x12" Exhibit - Art DC - Group Show - Wash DC - 2009

  • Robot Fest 09 - National Electronics Museum - Robotics Festival - 2009

  • DC Freshly Squeezed - Group Show - Wash DC - 2008

  • Ofrenda, Art of the Dead - Group Show - Wash DC - 2008

  • FLIK International Film Festival and Interactive Exhibition - Group Show - Wash DC - 2008

  • The Creative Process of Artist Groups - Gallery RFD - Group Show - Georgia - 2008

  • Gonzaga - Permanent Alumni Collection - Wash DC - 2007

  • Art Movements Past and Present - Gallery RFD - Group Show - Juror's Choice Award - Georgia - 2007

  • St Thomas  - Solo Exhibition - Wash DC - 2007

  • Splash Contemporary Art - Group Show - Ashbury Park NJ - 2007

  • The Gallaerie - Solo Exhibition - Wash DC - 2007

Featured In:

Learn more about Pindar Van Arman: