DescriptionAI technologies and their commercial applications in smart cities, personalized health care, and autonomous driving have developed rapidly in the last ten years. Innovators often present AI as the key technology to address grand challenges like climate change, health crises, and mobility. Yet, while the term “artificial intelligence” invokes ideas of disembodied algorithms, data, and cloud architectures, the material substrates of AI systems, made from minerals, fuel, and human labor, introduce some of the greatest challenges to developing AI in the public interest. While training neural networks based on modern graphic processors (GPUs) leaves a big carbon footprint, GPUs' production can rely on conflict minerals, accrue toxic waste, and currently requires massive amounts of water and energy. The project NeuroSys was launched in the spring of 2022 at the RWTH Aachen University, Germany, to develop more environmentally benign and ethically robust AI hardware. The project involves a transdisciplinary team of scientists, engineers, social scientists, ethicists, and industrial actors to develop neuromorphic hardware. By developing domain-specific neuromorphic hardware that integrates new materials with specific features summarized under the label “memristive,” the aim is to create more resource-efficient AI products. Moreover, the project approaches ethical questions, such as conflicting values in the production process and responsible applications of AI, as an integral part of technology development. In this presentation, Mareike Smolka will share some initial methodological proposals for engaging scientists, engineers, companies, and public stakeholders to study and co-shape the emerging innovation ecosystem around the production of neuromorphic AI hardware in the region of Aachen in Germany. For this purpose, she is interested in exploring scenario workshops, decision-making tools, and reflexive governance approaches. Drawing on her experience in Socio-Technical Integration Research (STIR) with neuroscientists, she will first discuss how collaboration between technoscientific experts and an “embedded humanist” can support responsible innovation in research and innovation projects like NeuroSys. She then invites the audience to reflect on the role of social science and humanities research in aligning innovation ecosystems with socio-ethical considerations.
|Period||5 Apr 2022|
|Event title||Public Interest Technology Colloquium Series|
|Location||Tempe, United States, Arizona|