Abstract
The proliferation of generative AI in the past two years raises complicated ethical and philosophical questions about the nature of AI systems, their potential, and the challenges and opportunities they bring about. The religious sphere is not immune from such concerns. In fact, it may be the arena where the status and role of AI unfolds in some of the most interesting ways.
In June 2022, months before the public release of ChatGPT and the ensuing generative AI revolution, a Google engineer started claiming that an AI program he was internally testing had become self-aware. His assessment was largely driven by his own religious beliefs and the AI’s predilection to discuss existential topics. It was an early demonstration of how the intersection between religion and AI would come to play an increasingly prominent role in public debates over the nature and ethical use of these technologies. If AI is to become ubiquitous, how will it be integrated into religious practices? If AI chatbots started to generate religious discourse and even claim to have developed religious interests and beliefs, how could we evaluate the authenticity of such outputs? Is it possible, at least theoretically, for artificial systems to develop anything close to what in humans we call religion?
To explore these questions raised by the nascent generative AI technology, theologian Marius Dorobantu convened an interdisciplinary workshop in Amsterdam in December 2022, supported by CLUE+, the interfaculty research institute at the Vrije Universiteit. Some of the articles in this thematic section, guest edited by Dorobantu, emerged out of that workshop, while a few others were subsequently added. The fundamental question explored in all of them, from multiple disciplinary angles, is whether AI could play a significant role in religious life, either as a tool in human religiosity or as an authentic religious subject itself.
Pilosopher Pim Haselager highlights how comparing human cognition with AI can significantly inform humanity’s journey of self-understanding, somewhat similar to the comparisons with animals and angels in medieval theological anthropology. He contrasts the smartness and cognitive abilities of AI with its complete lack of understanding and sentience, which he regards as sine qua non conditions of authentic religiosity. In another article, psychologist Fraser Watts and the late Yorick Wilks, AI pioneer, explore the feasibility and acceptability of AI-powered spiritual companions. Based on empirical research conducted with both GPT chatbots and Wizard-of-Oz methodology (humans masquerading as AIs), the article tentatively concludes that although artificial spiritual companions might proliferate, especially those that facilitate self-exploration, there are still dimensions of human spiritual counselling that might resist automation for the foreseeable future. Theologian Max Tretter intersects robotics, pop-culture and Christian eschatology to ask an intriguing question: is there an afterlife for robots and, if so, what might it be like? If there is hope for all creation to undergo eschatological completion, he argues, then the tentative answer to such a question could only be affirmative, opening up a discussion about the diverse conceptions of afterlife that can be imagined for robots. Computer scientist William Clocksin wrestles head-on with the question of whether intelligent robots could become religious. Following up on his 2023 article in this same journal, he makes a compelling case that future androids would likely develop a form of non-human personhood through sustained engagement in social relationships. If androids start questioning their place in the world and relationships with others, they might use religion in doing so, just like humans do, especially if they come to acknowledge their interdependence with others—human, robotic, or divine. Religious scholar Robert Geraci highlights the critical role religious beliefs and practices might play in recognizing artificial general intelligence, should it ever emerge. Reviewing the historical intersection between religion and robotics, the article critiques the colonial frameworks at work in evaluating “otherness”—of both humans and robots. It also reflects upon the need for an inclusive approach to integrating intelligent robots into our societies by acknowledging their potential for religious experience and participation. Theologian Daekyung Jung reaches a similar conclusion in his article, arguing from the perspective of embodied cognition. Future AI systems, particularly those integrated with soft robotics and driven by homeostasis as a fundamental goal, might exhibit religious behaviors if they develop human-level intelligence and self-awareness. Such religious behaviors would serve the AIs as cognitive mechanisms helpful in navigating existential challenges and the need to transcend finitude. In contrast with Clocksin, Geraci and Jung, Marius Dorobantu concludes in his own article that authentic religiosity is an unlikely development in robots, despite the theological openness to such scenarios. He argues that religion’s deeply embodied, social, and phenomenological underpinnings in humans may not be replicable in AI systems due to their fundamentally different bodies, cognitive architectures and needs.
Original language | English |
---|---|
Pages (from-to) | 661-787 |
Journal | Zygon: Journal of Religion and Science |
Volume | 59 |
Issue number | e |
Publication status | Published - 2024 |
Keywords
- artificial intelligence
- robotheology
- digital religion
- digital theology
- ethics of artificial intelligence
- Cognitive science
- philosophy of artificial intelligence
- science and religion
- theology and science
- theological anthropology
- artificial general intelligence
- ChatGPT