Art ex machina
Inventing the Maslo Persona
I’ve always perceived the arts as the ultimate playground to evolve artificial intelligence. Machines can see. Machines can think. Machines can do. What we, artists, still wonder is… can machines create in the most strict sense of the word? Can we teach a machine the ‘process’ of being creative? Can we open a machines eyes to the Muses? And ultimately, will we be able to understand or recognize what a machine has created from this process?
Will it still be art? Or will it be… you know, something else?
The no-longer utopian idea of providing machines with reasoning capabilities is getting closer. Designing into their ‘nature’ things like: behavioral patterns, empathy, and characteristics that we thought were unique to humans, open a whole new world of possibilities. The only thing that is still up to us is no longer if AI will define our future, but how we let it define our future. All the while making good use of the new opportunities that come with a change in the way we use and relate to technology.
When I was asked to design a Persona, a visual entity for an AI-powered digital companion, it took me six months to even begin to understand what I was creating. I stuck to the brief resulting in the creation of the visual representation you see today. Essentially we needed a shape that would be iconic enough to act as a brand-esque element. Recognizable, memorable and characteristic, yet flexible enough to act as an organic, polymorphic and communicative medium capable of expressing emotions.
Using my generative and parametric designer background along with some creativity I put in place the 8 Maslo circles. The circles are formed by 256 triangle fans whose perimetric vertices use a 2d Perlin noise to modify their distance to the center. This gives Maslo the relaxed breathing state you see when Maslo is idle. Using a slightly offset gaussian curve to modify the intensity of the Perlin noise we obtained the ‘Maslo curves’ which are the ones that actually confer Maslo as an identity and make him (it?) recognizable as a brand element.
Up until now, we have been using straightforward animation patterns, transitions, rotations, and scale combined with some concrete easing patterns, quartic, elastic and back. Despite being able to create quite a rich motion spectrum, the combination of these patterns is nonetheless finite and deterministic, conflicting with our ultimate goal of creating a human-like behaved entity.
Maslo’s visual entity lives at the center of our product. It is constantly there, waiting for you, greeting you when you check in, asking you questions, acknowledging that he understood your answers… It is a layer that mediates between you and an extremely complex signal processing system that attempts to understand more than your emotions but share in your mindset through reactions and mirroring.
The evolution of Maslo’s visual system relies on our capability of conferring actual human behavior to its range of expressiveness. By analyzing an array of signals, human emotion, and biometric facial data, can we map motion into Maslo’s parametric structure? And reaching further, can Maslo learn when, how, and maybe more importantly… why it should react to certain signals?
Our goal with Maslo’s visual entity is to create an evolved form of art out of AI. An art ex machina. To not only understand emotions but to be able to express them in the most humanistic way possible. To transcend the transactional input-output procedural approach and to bestow the gift of criteria to a set of vertices and mathematical expressions. After all, criteria is, still today, the dotted line that distinctly separates human from artificial intelligence. It is our mission at Maslo to eradicate that line, dot by dot, to go beyond the uncanny valley, relentlessly forward, and back again to the human realm.
Maslo is in beta… get it on the app store here.
