In times of constant busyness, technological overload and the demand for permanent receptivity to information, doing nothing is often seen as provocative and a waste of time. However, enjoying a moment of inaction and introspection while letting our minds wander and daydream may be more productive than staying constantly busy. In order to promote a doing-nothing state in 2019, Emanuel Gollob and his team created a neuro-reactive installation, using live EEG-measurements and a real-time adapting robotic choreography. Over time a generative algorithm increasingly learns to move the installation in a way that best supports the user’s mind-wandering process.

In order to promote a doing nothing state in 2019, Emanuel Gollob and his team created a neuroreactive installation, using live EEG-measurements and a real-time adapting robotic choreography. Over time the algorithm increasingly learns to move the installation in a way it best supports the user’s mind-wandering process, while a parametric real-time robot control system allows creating a space of more than 4 million possible choreographies of movement, sound & visuals. "Doing Nothing with AI" is an individually adapting robotic performance based on organic movement, peakless rhythmic narration and sine wave-like repetition - just catchy and interesting enough to help the user to get into a state of drifting thoughts while watching. A big part of the project is testing several aesthetics, all of them inspired by situations that still catch our eye today and could provide the potential of letting the perceiver’s mind wander. While in previous installations Emanuel Gollob explored the Doing Nothing potential of geometric motions, he and his team are now focusing on choreographies of more organic movements.

Every time a user puts on our EEG, one choreography is chosen based on already collected experience from previous users. At the same time, the EEG starts measuring the user‘s brain activity and after 30 seconds the current choreography gets evaluated. If it didn‘t bring the user closer to a state of drifting thoughts a GAN algorithm generates the next choreography. Using „Mx Automation“, a parametric real-time robot control system together with the Grasshopper plugin „Kuka PRC“ creates a space of more than 4 million possible choreographies of movement. For public interactive settings, I use a Muse EEG headband measuring the relative change of alpha and beta waves in the frontal brain. The EEG data is streamed wirelessly via Bluetooth. All positively evaluated choreographies are added to the training dataset and used to train a PyTorch GAN network. In the case of a negative EEG response, this GAN network generates the next choreography. The physical challenge was to create a material that translates large scale robotic movements to micro surface-motions. This particular version you can see here features 19.000 tricolored toothpicks regularly placed on foam multiplying the motion effect by changing the visual appearance during the smallest move in any direction.

Requirements: SPACE minimum 4 x 4 m with a minimum height of 3m Preferable exhibited in a space without sound distractions FLOOR If 4 anchors are possible, the leg construction can be skipped ELECTRICITY 1 light current slots SERVICE FROM YOUR SIDE WHILE SETTING UP Transport from Austria and back Travelling costs and accommodation for 1 artist SERVICE FROM YOUR SIDE WHILE ON SHOW Equipment Insurance Info person guiding the EEG interaction Hardware: „Doing Nothing with AI“ is animated by a Kuka Robot KR6 R900, a small-sized industrial robot with a weight of 50kg and the electrical power demand of normal 230V. Its height is 1.3 m, 2 m inclusive aesthetic skin, and its movement radius is 1m. As mentioned above, for public interactive settings, we use an easy to use Muse EEG headband. For more advanced settings we experiment with an Enobio EEG Cap from Neuroelectrics. In an ideal setting, we have a doing nothing space with a size of 4 x 4m and great lightning. SUMMING UP: Kuka KR6 R900, Kuka KRC4 Controller, Muse Headband 2016, 1x PC, 1x Sick Safety Scanner Software: Mx Automation, Rhinoceros, Grasshopper, Kuka PRC, Pytorch DCGAN, Python In general, all equipment is ready to be shipped.