The Cantor Center for Visual Arts at Stanford University came to us and proposed a collaboration with artist Trevor Paglen, whose work addresses topics like government secrecy and surveillance, exposing the vast apparatus of machines, systems and algorithms that monitor virtually every aspect of our lives. Paglen’s “Sight Machine” project would demonstrate to a live audience how machines “see” the world — in this case, a performance by the renowned Kronos Quartet.

Obscura worked with Paglen’s team to develop the computer and video systems to take a live video feed of the string quartet’s performance, run it through actual off-the-shelf artificial intelligence surveillance algorithms (over a dozen of them in total), and project what the AIs see and how they interpret it onto a screen above the musicians.

These AIs — whether for facial recognition, object identification or threat detection — are designed to communicate with their machine counterparts, not to provide human-readable output. Making that possible in realtime required Obscura’s systems engineers to maximize throughput in a Herculean research and development effort.