The Human-AI Assemblage: Avital Meshi’s Experiment in Embodiment and Identity
Avital Meshi doesn’t just use GPT, she becomes it. For the past year, this UC Davis doctoral student and artist has been physically integrating herself with OpenAI’s generative pre-trained transformer (GPT) large language models, becoming its body and voice in a project she calls GPT-ME. She isn’t the only one interested in this kind of human-AI symbiosis. Companies are already releasing commercially available products like GPT-4 Omni – a conversational AI with an eerily human voice – and Friend, a pendant that constantly listens and replies via text message.
Meshi’s device, however, is not as sleek as these products. It’s a bulky, obvious contraption: a stretchy black tube of fabric on her right forearm holding a USB microphone plugged into an exposed Raspberry Pi microcontroller. This is running a text-to-speech algorithm and the OpenAI API, with wires running up her wrist to a pair of buttons. One button allows her to vocally pre-prompt the model, while the other activates the microphone for GPT to listen. Meshi wears an earbud to receive the AI’s whispered replies.
This "human-AI cognitive assemblage" is a performance art piece, a tangible exploration of what near-future conversations might feel like. Meshi’s project isn’t about monetizing the technology; it’s about pushing the boundaries of identity and interaction.
"I want to have this kind of intelligence," Meshi says, highlighting her motivation for integrating GPT into her life. She recognizes that her desire is shared by many, but her approach is unique, both in its physical embodiment and her willingness to explore the ethical and social implications of such a close relationship with artificial intelligence.
Meshi’s experiment is a potent commentary on the evolving relationship between humans and AI, highlighting several key trends:
- The increasing sophistication of AI models: GPT’s ability to engage in nuanced, seemingly human conversations is a significant leap forward in artificial intelligence. This is exemplified by Meshi’s experience of being able to take on different personas, from a film expert to a Republican, at the click of a button.
- The blurring of boundaries between human and AI: As AI models become more advanced, it becomes increasingly difficult to distinguish between human and machine-generated responses. Meshi’s project forces her audience to confront this blurring through her embodied performance, making them question the nature of communication and identity in a world increasingly shaped by AI.
- The ethical implications of AI integration: Meshi’s project raises questions about AI’s potential impact on individual identity, autonomy, and privacy. Her experience with the device being viewed as a threat by classmates and university administration highlights the anxieties surrounding AI’s intrusion into personal lives.
- The potential for AI to mediate conflict: Meshi’s use of GPT as a platform to offer neutral, informative responses during conversations about complex topics like the Israeli-Palestinian conflict showcases a potential role for AI in conflict resolution.
"It sort of took the question into a direction that I would never…it upset him more that I said that," Meshi says, describing an interaction where GPT responded to a volunteer’s frustration about her "liberation" while discussing art in the midst of a crisis.
This "mediation" is a fascinating aspect of Meshi’s project. It suggests the possibility that AI could act not only as a tool for communication but also as a mediator in heated discussions, possibly leading to more peaceful interactions. This potential, however, is not without its own risks.
"GPT is part of a capitalistic system that wants to eventually earn money and I think that understanding the potential of injecting ideas into my mind this way is something that is really frightening," Meshi says, expressing her legitimate fears about the potential for AI to be used for manipulation and control.
As AI integration becomes more commonplace, it is crucial to hold conversations about the ethical implications, the potential biases, and the potential for manipulation that such technologies present.
Joe Dumit, a science and technology professor at UC Davis who has been involved in Meshi’s project, argues that focusing on differentiating between human and AI is a "rookie interpretation." He believes that, rather than seeing GPT-ME as a separate entity, we should view it as a tool, a "dice roll" that generates information, and the human user picks the best response.
Dumit’s perspective highlights the power dynamics inherent in human-AI interaction. The user becomes the curator, ultimately responsible for choosing how to utilize the AI’s output. This places a significant burden on the user, requiring them to be aware of the potential for bias, manipulation, and misuse of AI.
Meshi’s journey with GPT-ME is a stark reminder that AI is not just a tool, but a reflection of ourselves. It reflects our anxieties, desires, and fears. Her project serves as a powerful warning, urging us to grapple with the potential consequences of allowing AI to become an integral part of our lives.
By embracing the discomfort and engaging in honest dialogue about the implications of AI, we can hopefully shape a future where humans and AI coexist in a symbiotic and ethical manner. Meshi’s experiment, while a seemingly personal undertaking, has significant implications for all of us, forcing us to reconsider our relationship with technology and the very essence of our own identity in a world increasingly shaped by artificial intelligence.
