The world watched last September as millions mourned the death of Queen Elizabeth. Subsequently, many watched with great interest as Charles was enthroned as...
Author:
Author:
Author:
Author:
Author:
Author:
“An abba said, ‘The prophets wrote books, then came our fathers, who put them into practice. Those who came after them learnt them by heart. Then came the present generation, who have written them out and put them into their window seat without using them.” (The World of the Desert Fathers)
It seems appropriate that OpenAI launched its prototype ChatGPT (Generative Pre-Trained Transformer) at the end of November 2022. Its advent with initial reports of promising potential—and unsettling possibilities—began to circulate just as the frenzied final weeks of the semester geared up. As much as I wanted to experiment and see its capabilities for myself, producing (and codirecting) dress rehearsals for a children’s nativity play zapped the last of my bandwidth. So, until grades were posted and costumes packed away, I made do as we do in the twenty-first century and followed the conversations of friends and colleagues on social media. It was obvious from the stunning results being reported that a new tool with Promethean power would need to be reckoned with and sooner than expected.
For those unaware, ChatGPT is artificial intelligence “chatbox” technology that draws on large volumes of data amassed from the internet and other public sources. It has the ability to optimize vast stored banks of language models to generate clearly written, grammatically correct sentences for nearly any inquiry made of it. ChatGPT is designed to mimic human conversation and can recall long strings of dialogue without contradicting itself. In many ways, ChatGPT acts like enhanced digital assistants, such as Siri or Alexa, but is capable of more complex and demanding tasks in any number of academic fields: physics, mathematics, literature, and, of course, theology. Not only can it predict the path of a ball thrown across a baseball field, but it can write a cohesive sermon outline (or a sermon) based on any biblical text, or even, as one of my Facebook acquaintances requested, write a fairly detailed syllabus for a seminary class in biblical studies. When I could indulge in a little experimentation, it responded to my prompts to write a 250-word essay on the need for creative writing in theological studies, a 500-word devotional on the Christian need for silence, as well as a 1,000-word essay on how AI will benefit Christian pastors in their ministerial work. And because computers can retrieve information far faster than any human could, the astounding speed at which it operates makes any task it performs seem relatively effortless. Each task I asked it to perform, as well as the seminary syllabus mentioned above, was completed in less than half a minute.
As with any prototype being field tested, ChatGPT has its hiccups. While it writes with flawless grammar, the style is banal and sentences are often repetitive. Errors are inevitable and span the spectrum from being subtle to glaringly obvious. I found, just as others reported, when asked to document sources, mistakes increased. As AI technology, it is designed to learn from mistakes and improve performance with use. Still, the degree to which it can learn to distinguish truth from fiction from its “neural network” is concerning. AI’s ability to blend fact and fiction to create its own brand of misinformation is a serious red flag for society at large, never mind the integrity of theological education. The potential to undermine academic integrity through “AI-giarism” is being considered and a water-marking scheme might combat those who look to make illicit shortcuts. But, any measures currently discussed are pragmatic and technical solutions, not theologically informed responses.
From a theological perspective, AI will only ever be a tool of humanity. It is an imperfect creation designed by created beings who have already proved fallible. Without the grace-filled help of the Creator himself, humanity and every one of its creations are subject to its own futility. It may be incredibly fast and alarmingly efficient, but even ChatGPT “knows” it is as biased and discriminatory as the content it draws on. And, as a non-living entity with a soul, it cannot know the inspiration of the Holy Spirit. It is not sentient. The sermons and devotionals I asked it to create were reasonable and logical, but devoid of perspective born of lived experience and creative imagination.
The advent of technology as powerful as ChatGPT and its successors is worthy of the buzz it has generated. Any epiphany to be had is ours as a human race, not AI’s. The wisdom and witness of previous generations remind us that the arrival of life-altering, labor-saving devices isn’t new—only the technology is. As humans, we will learn to make use of and have respect for AI, just as our ancient ancestors did when they tamed fire, mastered metallurgy, and invented the printing press. The future might be here now. As a Christian community, with the help of prayer and discernment, we already have the tools to navigate its perils and promises.
Grimm, a John Wesley Fellow, is assistant professor of congregational formation, Wesley Seminary, Indiana Wesleyan University. She is an ordained deacon in full connection with the North Carolina Annual Conference of The UMC.