Embodied Artificial Intelligence – Possiblities and Controversies Born in the Value of Fantasy of the Body
“Embodiment” is probably one of the most important concepts in media studies, referring to the essential position of the body as an information-filtering medium in the interaction between the human and the external environment, which is crucial to the construction of human perception. In other words, the body is in the essence of the definition of being a human being. Yet technologies denaturalize our body. When the brain activity of Elon Musk’s “cyborg pig” is transmitted in real-time to a nearby screen through an implanted chip in the pig’s skull that collects signal with electrodes thinner than a human hair, the lines between machine and human is blurred as mechanical devices become part of our human body, monitoring, even shaping human behaviors and cognitions (Dvorsky 2020).
More radical is the portrait of artificial intelligence (AI) in many sci-fi novels and films where the program obtains human-like appearance and biomechanical bodies, and even self-consciousness, becoming indistinguishable from real human beings. Media theorist Paul Dourish once claimed that “embodiment is the common way in which we encounter physical and social reality in the everyday world” (Dourish 2004, 100), emphasizing the core concept of a body-based information-filtering and “encountering” process. Based on this explanation, the definition of human may expand to include cyborgs since they perform the same mechanism in information processing and construct their perception and understanding of the environment through filtered information based on mobility and “body-like” mediation. Yet androids are particularly important and different in this case. Essentially being AIs, the information processing mechanism for androids are at heart computational, filtering information through algorisms instead of “body-like” mediation, which allocates them outside of the definition of humans. Yet androids are not merely machines, either. In 2016, Google’s Go-playing AI – AlphaGo – played the historical move 37 in its second game with one of the world’s best Go players Lee Sedol, defeating the latter with this “move that no human ever would” (Metz 2016). Through deep learning about existed human masters’ moves and reinforcement learning where the AI played against itself for millions of games in its neural networks, AlphaGo managed to come up with new strategies of its own. In other words, androids have the potential to do beyond human command and generate new ideas. Therefore, different from cyborgs, androids might be seen as completely novel beings that are fundamentally different from humans or machines. They are a counterpart of human beings.
This is what makes it so important and exciting for me to think deeper about the history of AI, where the road leads to, and our relationship with androids in the future. After all, like it or not, we are already seeing prototypes of such intellectual beings – the famous human-like social robot Sophia, first developed by Hansen Robotics in 2015 and received citizenship in Saudi Arabia in 2017, which can communicate, read emotions, and have basic reactions when interacting with humans (Hanson Robotics, n.d.).
Paper Abstract
This paper analyzes the emergence of commercialized artificial intelligence (AI) assistant products and reassess the notion of the body, algorithmic culture, and human-technology relationship in the digital age. The 21st century has witnessed an expanding interest in AI assistants from fictional characters to industrial products. As the algorithm tends to humanize itself with “personality” or even a physical “body”, the discussion of embodied AI requires immediate attention for moral and ethical considerations. By tracing the genealogy of AI together with specific cultural and technological milestones such as the Hollywood film Her (2013), Amazon’s virtual assistant Alexa (2013), and Hanson Robotics’ humanoid Sophia (2016), this study questions the socio-cultural paradigms of AI, critiques the advertising campaigns by technology companies that seem to promote their products as being in some sense human, and further examines ethical dimensions of having such AI products in the personal realm such as the bedroom.
Project inspiration
- readings
- films
- videos
Paper draft (excerpt)
The development of artificial intelligence has come a long way through three stages, while the human-technology relationship also going through modes of external, commanding, and competitive, into the fourth and latest stage where the program is physically embodied just as human being. The inevitability of the latest stage lies in the symbolic significance of embodiment in culture and people’s long-time imagination and fantasy of the body in sci-fi novels and films. Sci-fi works inspire the practice in industry and laboratory, and such practice in return boost the prosperity of sci-fi works about embodiment. Understood this inevitability, the appearance of embodied androids suggests a possible intimacy between human and the machine at this stage, since according to affect theory, body build the ground for the development of authentic emotions. Meanwhile, this also brings up earnest and solid concerns about AI citizenship and ethics that indeed requires attention and further consideration. Yet since we cannot put the genie back to the bottle, probably the best way we can do now is to create an open discussion and educate more people and the next generation to recognize the problem and join the conversation. After all, solutions can and can only be found through experiments and practice.
Follow My Blog
Get new content delivered directly to your inbox.
Exploration of Media and Humanity 