In this essay, I argue that the concept of ‘trust’ is invalid in the case of our relationship with AI. I suggest the concept of ‘leap of faith’ of Kierkegaard as the better alternative.

From a Christian theological point of view, trust has theological connotations. We generally attribute the concept of trust to God, an omniscient entity. I do not know about God but God knows everything about me. Hence, I unconditionally trust in God. The concept of trust cannot be used in the realm of human beings. We do not have any reason to trust anyone. We do not know much about the people around us. Concept of trust came to forefront in western philosophy with Rene Descartes who claimed that the only thing that he is certain about is his existence. He doubted everything else. He later built a system of knowledge on that rock-solid foundation. Kierkegaard, the Danish existentialist argued that there is rupture at the bottom of reality and we do not have any reason to trust the human beings or the future. He recovers the fundamental uncertainty from the philosophy of Socrates. Keirkegaard’s idea of ‘leap of faith’ resurfaces in the existential phenomenology of Heidegger. It is found that the existential tradition denied the idea of trust in the level of human being and reality and emphasized the idea of ‘leap of faith.’

There is real problem when we introduce the idea of trust in human-AI relationship. Do we consider AI as a Divine being who is all knowing? Do we have any reason to have an unconditional trust in AI? AI seems to work on pattern recognition. It definitely gives us some meaningful output. Just as other human beings are black boxes for us, AI remains as another entity that cannot be comprehended. Hence, I argue that it is wrong to pre-suppose the idea of trust in human-AI relationship. I think that we can only have an approach of ‘leap of faith’ in the case of our relationship with AI.

Shibin Joseph, IIIT Delhi