Pentagram’s identity for Covariant’s AI software is centred around the process of learning
Working on a project regarding artificial intelligence (AI) requires being as adaptable as the technology itself. Pentagram’s recent identity for Covariant, a revolutionary AI software that allows robots to reason and act on their surroundings, is the perfect example of this.
“When creating identities for tech and AI companies, we are often trying to compress big and potentially complex ideas into a single expression,” explain Jody Hudson Powell and Luke Powell, the two Partners leading the project. “In Covariant’s case, we wanted to convey adaptability, learning and robotic intelligence.”
Obviously most people don’t have endless knowledge they can just reel off regarding complex robotics at this level, and the Pentagram team was no different: “We start with research and trying to understand the subject as deeply as we can. This process invariably leads us towards various essential technical and theoretical aspects of the project that lend themselves to an aesthetic solution.”
Their investigations eventually led them far outside the realm of design: “For inspiration, our research took us to a wide selection of topics ranging from; kinematics (the geometry of motion), neural net graphs and connected data points, flexible design systems and ultimately to decision boundaries (a visualisation of the inner workings of the deep artificial neural net).”
It is clear that their extensive research informed the centrepiece of the identity, the “Covariant Flow”, which they describe as “an abstract visualisation that represents the process of learning through what’s known as the ‘decision boundary’.” Complementing the AI software itself, this unique application adapts and changes in response to different inputs, altering the output image’s form, speed, complexity and colour. The end product can then be exported and utilised on whichever platform Covariant sees fit.