Romi Mikulinsky
Bezalel Academy of Arts and Design, Israel
Unlearning Machine Learning: New Narratives for Human-Robot Relations
The way we perceive of robots and the possibilities inherent in human-robot interaction is strongly related to historical and present-day narratives about intelligent machines. These cultural narratives are often dystopian, betraying human anxieties about loss of control, enslavement or making humans obsolete. The Frankenstein complex, as Isaac Asimov called it, has been central to the man-machine discourse.
While literary and cinematic robots are often humanoid, robotics itself is also inspired by the capacities of nonhuman animals. One famous example is Boston Dynamics’ non-humanoid robots that seem to spur anxiety about machine autonomy with their impressive hardware and robotic body movements. Whether humanoid or not, today’s most advanced robots process massive amounts of data gleaned from a combined network of sensors, machine learning (ML) and algorithmic processing to improve their navigation and interaction with humans and other machines. Parallel to ML becoming an interchangeable term with artificial intelligence, robots are increasingly integrated not only in factories but in retail, service, caretaking and other sectors.
The Silicon Valley techno-solutionism paradigm is shaping future society. Techno-solutionism or ‘technofixes’ refer to problem-solving technologies designed to solve a problem. As this logic spreads around industries and research, it is referred to more pejoratively implying an intellectual shortcut. In his radical 2018 “Manifesto on Algorithmic Humanitarianism,” Dan McQuillan writes that the nature of ML operations means they will actually deepen some humanitarian problematics and introduce new ones of their own. McQuillan calls for advancing emancipatory techniques and for creating human reasoning and additional criteria required to produce ML for the people.
What the mechanization of life might mean for human society usually remains very much in the background. In navigating hype and uncertainty, alternative narratives emanating from the arts could help counter public concerns. Moreover, new potentialities of nonhuman agents can be realized by going beyond the post-Fordist values promoted by the tech industry (f.e., productivity, efficiency, reliability and optimization and scalability). Opening up to additional examples of living with nonliving agents and learning new modes of learning, interaction and coexistence can enrich our communication with artificial intelligences. These new possibilities may also allay our fears and promote more holistic approaches resonating with humanitarian thinking and doing.
My paper examines how art and speculative design can blaze the trail for robotic agents and systems to operate in contact with diverse cultural norms and varied value sets. It poses new questions for the ongoing relationships human have with nonhuman agents and proposes alternative values for interaction with robots. What can we learn from humorous, generous, and even deliberately clumsy or useless robots? How could such traits contribute to more tolerant attitudes among humans and to more enlightened relations between human and nonhuman agents, and thereby contribute to society as a whole?