Computational ideation

Although AI is currently limited in its creativity, using machine learning algorithm as an ideation tool is slowly gathering traction. For example, there are open-source AI models with image-generation capabilities that use AI to design sculptures or create paintings that mimic great works of art. These capabilities aren’t just relevant to fine art, however, but have the potential to explore and test out new ideas and accelerate prototypes across design disciplines. Although new forms of algorithmically driven creativity are being developed, most of the AI field is focused on manually designing the building blocks of an intelligent machine, such as different types of neural network architectures and learning processes. But it’s unclear how these might eventually get bundled together into a general intelligence. Moreover, experts point out that teaching computers to be creative is inherently different from the way humans learn to create, although there’s still much we don’t yet know about our own creative methodology. Instead, others argue, more attention could be paid to AI that designs AI. That is, algorithms will design or evolve both the neural networks and the environments in which they learn by analogy with biological evolution.Yet, as suggested by IBM technologists, the goal is not to recreate the human mind but to develop the techniques of interacting with humans that inspire creativity in humans. That is, the augmentation of creativity, and how to get better efficiencies. So, AI can offer many benefits serving as a smart, efficient and inspirational assistant. Or, AI as an ideation tool.

Distributive ideation

Artificial Ideation, Aid, or the use of artificial intelligence to help generate ideas is in the early stages of development and dependent on further advancement in hardware and software programs, including programmable processors, open data (cloud storage, and historical data sets) and deep learning and artificial neural networks. Yet Aid is attractive in that it has the potential for offering a more efficient ideation process, and both in terms of range and speed of new ideas for innovation. But training artificial intelligence models in huge data centres, here ideas centres, might restrict or hold back Aid serving local needs or conditions (data discrimination or input bias). This suggests a decentralised model of Aid, or distributive ideation that would rely on decentralised rather than centralised computing power to facilitiate and encourage ideation at, say, smartphone level. Although distributive ideation would have less processing power than the hardware accelerators used in data centres, it would facilitate bottom-up rather than top-down ideation encouraging wider participation and collaboration in the design process. Also, distributive ideation would consume less energy and therefore have a positive impact on reducing carbon emissions. (This blog was triggered by researchers in University of Cambridge's Department of Computer Science and Technology set out to investigate more energy-efficient approaches to training AI models.)