ObjectResponder

ObjectResponder: Researching & Prototyping for Design collaboration with Artificial Intelligence

Malsattar, Nirav (TU Delft Industrial Design Engineering)

Giaccardi, Elisa (mentor)
van der Helm, Aadjan (mentor)

Delft University of Technology

Design for Interaction

2019-04-15

In this report, I have explained my iterative design process using research through design approach. This project focuses on the context of Artificial intelligence and design collaboration. It also represents a design method of integrating human and non – human biases while designing intelligent products.Abilities of Artificial Intelligence (AI) are expanding so rapidly, that it already surpasses the human in specific tasks that were not thought before. Recent advancements in machine learning algorithms (ML) and its techniques e.g. ‘deep learning’, enable the machine to develop creative content on its own (John, 2016). Meanwhile, in the design domain, people have already begun to consider artificial intelligence as new design material (Holmquist & Erik, 2017). One can consider it an intelligent design material as it can include creativity as individual machine learning models.To understand this new paradigm of using AI in the design process, I created a speculative prototype of a design toolkit called objectResponder (v1.0). A toolkit which enables to design and prototype from the perspective of AI in the ‘wild’(Rogers & Marshall, 2017). I explored this toolkit with six professional designers from various discipline. Initial results suggested that looking at the world from the perspective of the AI may enable designers to balance human and nonhuman biases, enrich a designer’s understanding of the context, and open up unexpected directions for idea generation. The results from the study initiate my graduation project with — identifying what designers need, their concerns and challenges while working with Artificial intelligence and Machine Learning projects. In my thorough investigation with professional designers and design students, I learned that there is a gap in comprehending Artificial intelligence technology in design practice. Such as, designers struggle to incorporate these technologies into their products and services due to the complex nature of it. It was also evident in the literature study that, designers’ need to understand the underpinning principle e.g. limitations of Artificial intelligence and Machine Learning (Dove, Halskov, Forlizzi, & Zimmerman, 2017). Designers’ currently working with Artificial intelligence technologies mentioned that they are looking for a tool or prototyping toolkit which integrates AI with embodied ideation and rapid prototyping methods.To understand the state-of-the-art of AI, a literature study was conducted with the exploration of various ML technologies and prototyping tools. The purpose of this literature study was to understand the state-of-the-art AI and its current state. In this literature studies, I encountered some initial prototypes of tools that showed the possibilities of Artificial intelligence intervening into the design process. Meanwhile, technology exploration with various AI and ML tools and platforms allowed me to learn some facet of current AI and ML tools and ML platforms and perceive its limitations. From this observation, I designed three varied computer vision enabled experiments. Designers from various expertise have participated in the experiments. They were asked to follow the idea generation process with and without an AI’s Computer Vision technique (Machine perspective). Based on designers’ feedback about the experience of working with a designed speculative prototype, I propose a design toolkit called ‘object responder v2.0’ — with further advancement in it.

Design for Interaction
Human-Computer Interaction
Artificial Intelligence and Design

http://resolver.tudelft.nl/uuid:dd33eba1-74e0-40c6-b1ad-a5a2aa92ce23

Student theses

master thesis

© 2019 Nirav Malsattar