How much of our lives should we be prepared to outsource to AI?
The Financial Times (FT) Magazine (December 17th, 2022) featured a fascinating article on Generative AI, written by Tim Bradshaw. A great deal of focus has been recently placed on Chatbot GPT, and the FT reported on the technology’s ability to create marketing strategies, write reports, and generate software code. However, the FT observed, the results are not always reliable unless the chatbot is provided with effective prompts. This is the case for all AI within this generative space, including image creation AI such as Midjourney. A quote within the FT article from AI entrepreneur Colin Treseler, summed up the current situation, the quote is paraphrased as follows- you need to find an effective way of talking to the technology to get effective results.
The above technological circumstances have given rise to a new job, Prompt Engineer. Essentially, a Prompt Engineer is someone who can ask Generative AI effective questions, a bridge between our world and the world of the technology. The article bought to mind the work of philosopher, Thomas Nagel.
Nagel (1974) is his seminal article, What is it Like to be a Bat? discussed subjectiveobjective viewpoint opposition. Crudely put, Nagel observed that we could know many things about the creature, the bat, such as how it navigates the external environment and its day to day and seasonal habits, but we can never know what it is like to be a bat. In other words, there is always a piece of data missing, the perspective of the object being observed. This leaves all knowledge with a sense of mystery attached to it.
Returning to generative AI, we could argue that the technology requires very specific or expert prompts because it does not have the perspective of a human being, it does not have lived experience. Therefore, to engage effectively with an AI, we need to provide it with perspective by communicating with it in a certain way. This is currently the developing domain of the Prompt Engineer. Since the AI cannot take a human perspective, its output is inconsistent since it cannot engage with the mystery involved in non-specific communication.
The mystery which comes with life has long intrigued and troubled human beings. Shamans and Oracles used to act as metaphorical bridges to the gods and nature to provide us with insight and create a new more rewarding perspective (Vervaeke et al, 2017). Rituals emerged as a way for whole communities to engage with nature and their deities with the aim of gaining insight into their lives both personal and collective (Han, 2020). Rituals were essentially a way of communicating with mystery in an ongoing transformative way. At its root, ritual is an appreciation of one of Nagels’s (1974) key points, there is always data missing in the observed.
The role of the Prompt Engineer in this instance is tied up to a degree with the role played by AI, technology, and knowledge in Western culture. When Han (2020) observed the decline of ritual and the consequences of this decline he argued the decline has corresponded with the rise in productive time over transformative time.
Transformative time is where we attempt to transform ourselves by gaining a new aspect on our lives and communities. Ritual and the accompanying symbolism (religious artifacts, jewelry, totems for example) invite us to engage with the mystery which comes with the missing data identified by Nagel (ibid). Time is spent engaging with ritual to facilitate what Wittgenstein (2010) called a “noticing of aspect”. In the beginning is the deed (Ibid drawing from Goethe) provided by ritual and the deed provides a new aspect; our perspective is transformed.
Our attempt to increase productive time by seeking knowledge to replace any mystery has naturally led to a decline in ritual (Han, 2020). The missing data present in all observations can be outsourced to science rather than contemplated in a process of transformation. This project has to a large degree failed, with McGilchrist (2019) arguing that our refusal to contemplate mystery and nuance and seek only certainty has skyrocketed clinical presentations of anxiety, depression, and schizophrenia. Once again, returning to Nagel (1986) it is not possible to account for everything objectively and scientifically, unless of course we can become a bat. The scientific view is the view from nowhere (ibid).
AI has offered another opportunity to restore the project of absolute knowledge. However, because our current AI developments cannot take the perspective of human beings, they require a bridge between our world and the digital. If AI could at some point take the perspective of a human being, then we may be replaced. However, would that rob us completely of transformative time? If we can outsource mystery to AI, would we be lost only in productive time whether that be leisure or work?
It raises some intriguing questions. Our decline in rituals has blunted our communication with mystery and our ability to engage in “noticing of aspect”. However, how far do we want to go? What aspects of our lives should retain mystery and seek to engage with mystery more deeply as opposed to erase it? To return to the comments made by Colin Treseler earlier in this article, about finding effective ways of communicating with AI, perhaps we also need to restore effective ways of acknowledging and engaging with mystery.
Reading
B. C. Han (2020) The Disappearance of Rituals: A Topology of the Present. Polity
Nagel, T (1986) The View from Nowhere. Oxford University Press.
Nagel, T., (1974) What is it like to be a bat?. The philosophical review, 83(4), pp.435-450.
Vervaeke, J., Mastropietro, C. and Miscevic, F., (2017) Zombies in Western Culture: A Twenty-First Century Crisis. Open Book Publishers.
Wittgenstein, L., (2010) Philosophical investigations. John Wiley & Sons.
McGilchrist, I., (2019) The master and his emissary: The divided brain and the making of the western world. Yale University Press.
