Shall the Luddites Win?—A Friendly Response on ChatGPT

I expected my article on ChatGPT to cause a stir, but I still wasn’t fully prepared for what I had set myself up for. In response to a number of ways I articulated students and pastors can use ChatGPT, my friend Jordan Steffaniak published an article in response. He argued that the dangers inherent in ChatGPT outweigh any perceived benefits and necessitate leaving ChatGPT alone.

In response to Steffaniak’s emphatic “No!” I’d like to offer a resounding “Yes!”[1] But before we can look at Jordan’s article, we first need to take step back to address the necessary theological framework to approach cultural issues by drawing on the work of Andy Crouch.

A Theology of Cultural Engagement

In his helpful book Culture Making, Andy Crouch provides four common approaches to culture: Condemning, Critiquing, Copying, and Consuming Culture. It seems that Jordan’s self-professed luddite tendencies betray marks of the Critiquing Culture model.

In the Critiquing Culture model, engagement with culture is understood as “critiquing them carefully to show how they are inadequate or misguided.”[2] However, little else is done in terms of positive engagement with the culture. Jordan’s article doesn’t offer “cautions” or “challenges” with using ChatGPT. Instead, he simply offers problems with ChatGPT, contending the “latent nefarious potential [of ChatGPT] is so potent and corrupting that while tools like ChatGPT can theoretically be employed harmlessly, even beneficially, the risk is not worth the reward.”

Crouch points out the obvious problem when one only critiques culture: “Critique and analysis very rarely change culture at all.”[3] To be sure, thoughtfully investigating cultural artifacts like ChatGPT must occur when interacting with culture, but merely boycotting ChatGPT will no more change the culture than railing at a fiddling Nero would have put out the fire in Rome. We have to offer more than an emphatic “No!”

Instead, Crouch proposes the idea of cultivation in which “people…tend and nourish what is best in human culture…[and] do the hard and painstaking work to preserve the best of what people before us have done.”[4] Of course, there is room to critique in this model, but Crouch argues “those who have cultivated and created are precisely the ones who have the legitimacy to condemn—whose denunciations, rare and carefully chosen, carry outsize weight.”[5] I propose that Crouch has cast the right vision for cultural engagement. This model of Cultivating Culture should be the posture that we should employ when confronted with cultural artifacts like ChatGPT.

And it does seem that we can approach ChatGPT with a posture of cultivation, as there is nothing inherently problematic to ChatGPT. Even Jordan avers (all the while disagreeing with the use of ChatGPT), “While tools like ChatGPT can theoretically be employed harmlessly, even beneficially, the risk is not worth the reward” (emphasis mine). In other words, his problems with ChatGPT are not inherent to AI. ChatGPT is a cultural artifact that can be cultivated. Indeed, in Crouch’s analysis, ChatGPT must be cultivated.

Putting Theology to Work

With a posture of cultivation towards ChatGPT in place, we can begin to respond to Jordan’s objections that ChatGPT can “make us deeply irrational, slaves to vice, and ultimately unhuman.” Our framework of cultivation would have us see these objections as challenges, as barriers to overcome in employing ChatGPT for faithful stewardship.

Indeed, we must not merely cower in fear at the possibility of such drastic frightful allegations. We must do the calculus to consider the probability that ChatGPT will have such deleterious effects. In much the same way that we weigh the risks and rewards of air travel (which can end in death!), we must also weigh the chances that the risks actually materialize.

Jordan argues that “ChatGPT teaches us to be irrational.” As we use “ChatGPT with a degree of trust that isn’t earned” and “attempt to create a machine” to replace God-given rationality, we will fall into “us to slavish dependence that stifles rational thinking.” But it seems that Jordan’s warnings provide the necessary antidote to keep us from this slavish dependence. If ChatGPT really does have such limitations, then wouldn’t knowing about such limitations force us to fill in the creative gaps? If ChatGPT cannot be depended on for rationality, what would keep us from simply using ChatGPT for “more informed decisions by presenting data-driven insights and suggesting logical connections between concepts,” as ChatGPT itself suggests?

Jordan argues that “ChatGPT teaches vice instead of virtue” by offering a shortcut to pass off work as your own. In reality, ChatGPT is simply the latest development in the arms war between students and teachers. Ripping off commentaries and plagiarizing sermons began long before overreliance on ChatGPT. The problem here isn’t technology; it’s humans. The solution isn’t trying to temptation-proof our life but to be matured to resist temptation. While a Critiquing Culture mindset would suggest wholesale abandonment, a Cultivating Culture mindset would urge teachers to form students and pastors spiritually as well as intellectually so that they are equipped to rightly use technology to its full potential. Instead of simply warning of the dangers of ChatGPT, we must give ways for students and pastors to properly utilize this technology.

Lastly, Jordan contends ChatGPT “can destroy our ability to be human” by robbing us of an intellectual community and undermining our finitude which is at the heart of our humanity.[6] But again, Jordan simply provides helpful considerations if we were to cultivate ChatGPT. ChatGPT can never replace intellectual communities, but it surely can provide helpful idea generation and feedback. And what else teaches us dependence other than using tool like ChatGPT?

There is no retreat from ChatGPT. AI is not disappearing anytime soon. We eventually must develop an interactive theology of AI. We can either chose from a Critiquing Culture mindset and have nothing positive to offer or the Cultivating Culture posture and provide a robust model for Christian engagement. To be sure, the challenges that Jordan has mentioned will not be the last. But viewing these problems as challenges to overcome will aid Christians in developing a robust theology of technology to interact with ChatGPT.[7]

References

[1] This echoes Barth’s reply to Emil Brunner on natural theology entitled “No! A Response to Emil Brunner.” See https://www.britannica.com/topic/Nein-Antwort-an-Emil-Brunner.

[2] Crouch, Culture Making, 68.

[3] Ibid., 68.

[4] Ibid., 97.

[5] Ibid., 98.

[6] I’m not entirely sure that finitude is as much essential as it is common to humanity. Thomas Morris argues that the essential properties of humans are being rational, moral, spiritual, and aesthetic while finitidue is merely a property that is common among mankind. Christ, therefore, was able to be fully human (because he was rational, moral, spiritual, and aesthetic) while still being fully God. See Thomas Morris, The Logic of God Incarnate (Cornell Univ. Press, 1986).

[7] Thanks to Dennis Nicholson for his helpful feedback on the article. And I’m indebted to ChatGPT for aid with stylistic variety.