How AI could change our relationship with religion
Published: 5 March 2024
Research insight
Prof Sreevas Sahasranamam explains how science and faith are often kept in two distinct boxes that hardly ever intersect. However, he believes that as AI becomes more mainstream, it will fundamentally alter our engagement with faith and spirituality.
Looking at the ongoing developments, Prof Sreevas Sahasranamam observes a significant shift in the landscape. Many ancient faith-related texts, originally recorded on papyrus and palm leaves, face accessibility challenges in the modern era, primarily due to two major obstacles.
Firstly, a considerable portion of these ancient texts still in existence are fragmented, susceptible to deterioration over time. Secondly, even those texts that have been digitised present a barrier as they are often written in languages understood by only a handful of individuals today, rendering them inaccessible to the majority.
However, Prof Sreevas Sahasranamam notes that AI is fundamentally reshaping this scenario by facilitating easier access to these texts. A notable instance from 2023 showcased how computer scientists at the University of Kentucky utilised AI to unveil the contents of carbonised papyrus that had been charred during the eruption of Mount Vesuvius in AD79
Examining 3D X-ray images of the papyrus, scientists directed their focus towards training AI algorithms to decipher letters within the scrolls. Leveraging subtle structural cues left by ancient ink, the AI successfully interpreted and translated the ancient Greek term for "purple" inscribed on the scroll. This breakthrough is particularly significant given the linguistic complexities present in other ancient texts, such as those from the Indian subcontinent, which are often in languages like Sanskrit or scripts like Grantha, understood by only a select few today.
In parallel, AI systems are harnessing language translation and predictive text capabilities to anticipate missing characters in ancient inscriptions. Start-ups are pioneering the use of small language models trained on Sanskrit texts, employing them to develop AI-guides—referred to as AI-Guru or AI-tutors—to aid individuals in navigating these intricate texts. Furthermore, AI's potential extends beyond textual preservation, as it also demonstrates promise in the restoration of faded cave paintings.
Like any other technological advancement, AI presents its own set of challenges, particularly concerning misinformation and misrepresentation within the realm of faith. A notable example from last year involved the viral circulation of a deepfake image of the pope, initially intended as a playful prank. However, the ease and affordability of creating deepfake images and voice clones through AI, some costing as little as £10 using readily available online tools, raise concerns about more sinister applications. Consider the possibility of extremist groups utilizing this technology to fabricate propaganda videos featuring the falsified words of religious leaders or distorting religious symbols for malicious purposes.
Moreover, AI language models trained on extremist ideologies could generate large-scale outputs promoting hatred against specific faith communities, potentially inciting hate crimes and exacerbating community tensions.
Transitioning to a futuristic perspective of AI's intersection with faith, let's envision a hypothetical scenario rooted in the Hindu faith over the next 50 years. Central to Hinduism is the concept of purushartha, which delineates the four aims of human life: dharma (righteousness), artha (prosperity), kama (pleasure), and moksha (spiritual liberation).
In the forthcoming 50–100 years, AI is poised to become an indispensable co-pilot in all human endeavors, including decision-making processes. Upholding righteousness or dharma in society would necessitate training AI algorithms and language models on data that is not only accurate but also morally sound. This imperative underscores the importance of integrating ethical considerations into the regulatory frameworks currently under development for AI.
Advancements in AI are anticipated to propel humanity towards economic prosperity (artha), ushering in an era of technological abundance and convergence. Addressing the escalating issue of societal inequality may lead to the widespread adoption of initiatives like universal basic income, affording individuals more leisure time to pursue activities beyond their primary occupations, including pleasures (kama).
With an abundance of free time, there may be a resurgence of interest in spiritual and faith-related pursuits (moksha). This potential revival underscores the profound influence of AI on reshaping societal dynamics and human aspirations, promising a future where technology and spirituality converge in unexpected ways.
Read the full blog on The Conversation webpage.
First published: 5 March 2024