Speculative biosphere - an experience report on the innovation camp ‘Gestaltungsmaschine’

How will artificial intelligence (AI) shape the economy and society of tomorrow? What does AI mean for the interaction between man and machine? And what does AI have to do with collaborative creative processes?

These questions were the focus of the three-day innovation camp ‘Gestaltungsmaschine’ of the Initiative Kultur- & Kreativwirtschaft of the German Federal Government on the last weekend of March.

With the beta version of our research portal in our luggage, we offered a workshop on the topic of ‘biodiversity, knowledge transfer and AI’ for the participating entrepreneurs, experts and stakeholders of the creative industries. Co-director of the workshop was the artist Sascha Pohflepp, whose work focuses on the influence of technology on the relationship between humans and nature.

Sascha also drew our attention to so-called CycleGANs – “The coolest idea in deep learning in the last 20 years.”. These networks proved useful to us later on.

In three workshop blocks, we discussed the significance of museum collections and data for the history and future of nature, as well as future uses in the context of biodiversity using artificial intelligence. Furthermore, we developed a speculatively designed use case out of this.

After an initial examination of the content of the Museum für Naturkunde Berlin and the data that will be available in the research portal in the future, the workshop dealt in depth with possible concepts of use that could be derived from the Mediasphere database for the future of nature. In the second workshop block, the group of participants chose the approach of speculative design for the development of alternative future ideas with reference to the present. This means the speculative further thinking of natural history based on past and present interactions between man and nature. A central impulse came from an artistic project about a moth – the peppered moth – and the connected concept of industrial melanism, which Sascha presented to us.

There are two morphs of the peppered moth, which differ in their wing colours: the light form has white wings with a dark pattern, the dark form mainly black wings. Traditionally, the light form ( Biston betularia) was more common in Europe, because it was better camouflaged on the bark of birches covered with light lichens, whereas the dark peppered moth ( Biston betularia f. carbonaria) was easy to detect by its predators, the birds.

The term industrial melanism describes the phenomenon of the indirectly industrially enhanced deposition of dark pigments in skin cells. The term originated at the end of the 19th century, when the frequency ratio of the morph of the peppered moth in English industrial areas shifted and mainly individuals of the dark phenotype were observed. Although the difference between light and dark peppered moth is genetically determined, the shift in the frequency ratio of the two variants is due to the decline of light lichens on the birch bark because of the accumulation of soot particles caused by industrial air pollution. Because of this change in the birch bark, the camouflage possibilities and also the chances of survival of the two phenotypes were reversed, so that the dark peppered moth had an advantage and appeared correspondingly more frequently [1].

Our team took this example of a human-technically influenced process of evolution as a starting point for the innovative idea generation: Which assimilation processes are conceivable for the future? And how can the digital data of the MfN and AI influence them? Therefore the concept of an AI was created, which develops and predicts new structures based on past and existing patterns. For this purpose, the AI would have to be trained with data on patterns of nature, for example from our research portal, on the one hand, and with data on patterns of the human or urban environment on the other.

As an example, our design of the ‘Berlin moth’ should illustrate the speculative design concept in a double sense:

Is it possible to think of assimilation processes in the human-urban environment through the processing of urban and natural patterns by AI? And what would such an ‘inverted’ adaptation look like?

Here are our results:

In the third workshop block, we finally discussed our results and prepared the presentation of our concept to the other 10 workshops. Beforehand, Sascha had told us about how he successfully got a CycleGAN running overnight. We learned a lot and are curious where this idea will take us.

References:

[1] Inge Kronberg (2016): Industriemelanismus: Die dunkle Form des Birkenspanners entsteht durch ein springendes Gen. In: Biologie unserer Zeit. Jg. 5, Heft 46, S. 276-277. https://doi.org/10.1002/biuz.201690083