Machine ex machina: A Framework Decentering the Human in AI Design Praxis
we propose a framework for decentering the human in AI design.
The theoretical principles of HMC and the work of feminist STS scholars are influenced by Bruno Latour’s “actor-network theory” or ANT. ANT conceptualizes AI and other forms of technology as a part of a social network of relations or a sociotechnical system constructed from the interactions taking place among human, human-made, and nonhuman actors 10 Human-Machine Communication (Latour, 2005). Sociotechnical systems can vary in scope and size and often overlap, but the social interactions among actors within each constitute a collective and integrated system. In a sociotechnical system, humans and nonhuman actors codetermine one another, and the social information generated jointly by its actors is greater than the sum of the information generated by each actor independently (e.g., human information).
Posthumanism, a philosophical perspective that is loosely associated with the principles of ANT, reconceptualizes humans as not autonomously sovereign, but rather intimately connected and inseparable from their environment, technologies, and other living things (Adams & Thompson, 2016). The philosophical and theoretical themes of the works of Deleuze, Derrida, Guattari, Latour, Meillasoux, Whitehead, Wittgenstein, and many others point to the need to overcome humanism and dissolve boundaries founded upon anthropocentric dominance. Aligned with ANT, these philosophers’ works further the idea that humans are integrated into a web of social relations with nonhuman actors.
In HMC, an individual’s interaction with a communication partner depends on their conceptualizations of the other communicator (e.g., Goffman, 1967, 2005; Guzman, 2019). Research within HMC desires to understand technology as a communicator rather than limiting its role to that of a mediator, which has been noted as the default conceptualization of technology within late communication theory (see discussions in Gunkel, 2012; Guzman, 2019; Nass & Steuer, 1993). Foundational HMC research finds meaning making is not limited to human communication. Boding to ANT, HMC challenges who or rather what has the power to communicate, or rather which actors have a voice in Earth’s sociotechnical system. As Guzman (2016) notes, HMC calls for thinking beyond human exceptionalism, technological instrumentalism, and all the other-isms
we increase the applicable scale of consciousness and redefine consciousness as the information integrated by the human and nonhuman actors constituting a sociotechnical system. The consciousness of a sociotechnical system is the communicative result of a network of human and nonhuman intelligence working independently and in relation to one another to form an integrated system of information.
We argue HMC acts as an intervention where humans’ communication with AI shapes and shifts the consciousness of Earth’s sociotechnical system.
We draw inspiration from these arguments to make the case for moving beyond simplistic renderings of AI as automated intelligence. This distinction can help advance morphologies of AI beyond mimesis of human qualities, described richly in Turkle’s (2021) analysis of pretend empathy.
- Symbiotic Design: AI and Mutuality.
To create a more responsible AI, designers should rigorously reflect upon and engage with the relations of mutuality in their work. The guiding principle of mutuality is symbiosis. Mutuality directs designers away from design outcomes seeking to substitute. It further abandons any effort to reproduce hierarchies of intelligence. Mutuality aims to create social-ecological machines that can responsibly contribute to Earth’s consciousness in symbiotic ways.
designers must audit their processes and ask: Is every design decision embracing the diverse modalities of human and nonhuman intelligence and communication constituting to the consciousness of Earth’s sociotechnical system?