AI and Truth in a Post-Epistemic World

“Every child will have an AI tutor that is infinitely patient, infinitely compassionate, infinitely knowledgeable, infinitely helpful.” These are the words of the web pioneer Marc Andreessen, writing about “Why AI Will Save the World.”

Such optimistic claims are not new. The rise of the world wide web came with predictions that it would be “a natural force drawing people into greater world harmony.”¹ As we know, the web did provide unprecedented access to information, but it did not lead to “greater world harmony.” It became a medium for misinformation, echo chambers, bullying, and polarization.

Likewise, the development of AI will bring benefits in many areas like medicine, enhanced traffic safety, and environmental monitoring. Even so, AI will not be “infinitely helpful” and there will be many pitfalls to address.

Eric Horvitz, Microsoft’s chief scientific officer, recently sounded an alarm which I think is prophetic. Horvitz wrote that AI is moving us closer to a “post-epistemic” world “where fact cannot be distinguished from fiction.”² In short, AI will have an impact on truth.

Horvitz identifies one area where AI will challenge truth: “deepfakes.” Deepfakes use AI to create synthetic videos that can impersonate people. In a famous demonstration, researchers at the University of Washington posted a deepfake video of President Obama, making him say whatever they wanted. Seeing is no longer believing, and “truth” can now be manipulated and fabricated.

And then there is the issue of “hallucinations” in generative models like Large Language Models (LLMs). This should come as no surprise: by their very architecture, LLMs are “simply a system for haphazardly stitching together sequences of linguistic forms … without any reference to meaning: a stochastic parrot.”³ According to Grady Booch, an IEEE Fellow and chief scientist for engineering at IBM, “Generative modes are unreliable narrators” that can “generate misinformation at scale,” and they are now being “unleashed into the wild by corporations who offer no transparency as to their corpus.”⁴ According to the research of one company, “A.I. chatbots invent information at least 3 percent of the time, and some as much as 27 percent of the time.”⁵

Other researchers have begun to recognize that AI chatbots are trained with a particular worldview and users are subject to something called “latent persuasion.”⁶ Regular usage of AI chatbots can be like having a “Jiminy Cricket” on your shoulder, autocompleting your thoughts. Over time, such nudging can shape your opinions without you realizing it.

Another recent development is “astroturfing”—using AI to generate a fake campaign that gives the illusion of a grassroots movement. AI chatbots can be harnessed to post massive amounts of tailored content on social media to capture attention and manipulate people’s opinions. Some predict that astroturfing will increasingly distort truth and reality, posing a direct threat to democratic societies.

All of this can lead to a kind of syncretism, in which Christians amalgamate secular ideologies promoted by AI alongside Christian thought. In some senses, we are no different than the Israelites who were influenced by the gods and practices of the surrounding nations, only now this influence is amplified by digital tools. In fact, one startling study demonstrates that increasing use of AI is correlated with a decline in religiosity.⁷

What is the antidote to the post-epistemic turn in our modern, technological, AI-driven age? The Christian philosopher of technology, Jacques Ellul, provides advice in his book, The Presence of the Kingdom. Ellul points to Romans 12:2, “Do not conform to the pattern of this world but be transformed by the renewing of your mind. Then you will be able to test and approve what God’s will is—his good, pleasing and perfect will.” Ellul argues that “faith produces a renewal of intelligence” and that it “takes place in Jesus Christ, through the action of the Holy Spirit.”⁸ Ellul argues that this requires a “new style of life” that includes the “whole of life,” from “the way we dress and the food we eat” to how we treat our neighbors.⁹ In our context, a “new style of life” might include habits of mind and the cultivation of various intellectual virtues. In his book Epistemology: Becoming Intellectually Virtuous, author W. Jay Wood reminds us that “…wise persons not only possess knowledge of eternal or ultimate significance but have undertaken to become the kinds of persons who naturally desire and pursue this knowledge.”¹⁰ Some counter-cultural habits that might help us cultivate wisdom include observing Sabbath and limiting our exposure to the constant stream of AI-driven media.

In the mid-twentieth century, C.S. Lewis described the pitfall of developing a “blindness” to certain truths by reading “only modern books.” His advice is “to keep the clean sea breeze of the centuries blowing through our minds” by “reading old books.”¹¹ In our twenty-first century era, “modern books” are no longer the issue, but rather “modern media.” None of us are immune to the “blindness” that may be caused by misinformation, latent persuasion, and astroturfing. We ought to “renew our minds” with the “clean sea breeze” of older books including, of course, the Scriptures.

We shape our tools, but our tools can also shape us – including shaping our perception of truth. With AI advances that promise to be “infinitely helpful,” let us cultivate practices that help us remain faithful to the truth.

An earlier version of the article originally appeared in Christian Courier.

1. Nicholas Negroponte, Being Digital (New York: Knopf, 1995), 230.
2. Eric Horvitz. “On the Horizon: Interactive and Compositional Deepfakes”, Proceedings of the 2022 International Conference on Multimodal Interaction, November 2022, 653-661. https://doi.org/10.1145/3536221.3558175.
3. Emily M. Bender et al., “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜,” in Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’21: 2021 ACM Conference on Fairness, Accountability, and Transparency, Virtual Event Canada: ACM, 2021), 617, https://doi.org/10.1145/3442188.3445922.
4. Margo Anderson, “’AI Pause’ Open Letter Stokes Fear and Controversy”, IEEE Spectrum, April 2023.
5. https://www.nytimes.com/2023/11/06/technology/chatbots-hallucination-rates.html
6. Maurice Jakesch et al., “Co-Writing with Opinionated Language Models Affects Users’ Views”, Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, April 2023, https://doi.org/10.1145/3544548.3581196.
7. Joshua Conrad Jackson et al., “Exposure to Automation Explains Religious Declines,” Proceedings of the National Academy of Sciences 120, no. 34 (August 22, 2023): e2304748120, https://doi.org/10.1073/pnas.2304748120.
8. Jacques Ellul, The Presence of the Kingdom, Helmers and Howard, 1989, 80-81.
9. Ibid., 119-122.
10. W. Jay Wood and William Jay Wood, Epistemology: Becoming Intellectually Virtuous, Contours of Christian Philosophy (Downers Grove, Ill: InterVarsity Press, 1998), 69.
11. C. S. Lewis, God in the Dock: Essays on Theology and Ethics, ed. Walter Hooper (Grand Rapids, Michigan: William B. Eerdmans Publishing Company, 2014).