Amy Isham
25 November 2024
An experimental podcast uses artificial intelligence to bring famous writers and philosophers “back to life.” But just because we can, does that mean we should? ISCAST supporter Amy Isham breaks down the implications.
A few weeks ago, a friend sent me a cultural engagement podcast he created using artificial intelligence prompts. The episodes are short and punchy, some as short as five minutes. Some are narrated like a news reader in a smooth, male, American accent, others have two speakers riffing in a believable conversation.
The podcast was designed to summarise top news stories from the perspective of great thinkers like C. S. Lewis, G. K. Chesterton, and Augustine. The creator experimented using a Python script, but since the release of Google’s NotebookLM, the process is much easier, so we’ll soon see many more like it.
Lewis, Chesterton and Sayers
As I browsed the list of episodes, I was drawn to the ones “by” C. S. Lewis, G. K. Chesterton, and Dorothy Sayers. These fascinating authors were known for their public Christian faith and for writing both fiction and nonfiction. This “triple-threat” ability to write insightful nonfiction, engage with culture, and produce sparkling and compelling stories is likely why their work continues to feel so meaningful.
Lewis is still widely known for writing nonfiction books and essays, fantasy children’s fiction set in Narnia, and speculative science fiction for adults. Lewis’s Screwtape Letters, short supernatural fiction about demons, was read on BBC Radio as late as 2013, a testament to his continued influence on Western culture.
Read more: Should we fear AI? A biblical perspective
Chesterton may be less known for his apologetics and philosophy, but he inspired Lewis, who loved reading his Father Brown detective stories. Sayers is known for being one of the first women to graduate from Oxford University, as well as her essay Are Women Human about female identity, but she’s also known for her television-adapted detective novels about Lord Peter Whimsy.
My first reaction to discovering the synthetic podcast was unease. My long-dead heroes were speaking from beyond the grave in human-sounding voices, on issues such as the war in Gaza, the regulation of the gambling industry, and the latest American election craziness. If AI can write and narrate content faster than humans, where will these products be in a few years? As someone who may take weeks to plan, record, and edit their podcast episodes, was I out of a job? Is there value in continuing to craft human-generated content when synthetic neural networks can do it faster?
When I recovered from my horror I was seized with a powerful thirst for more of these short missives. I craved guidance from wise and trusted thinkers on the confusing, baffling, and overwhelming issues of today. I was desperate to know what John Frame, the now-retired Christian philosopher, thought about the proposed social media ban for teenagers in South Australia. I imagined what it would be like to read a sensitive and heartfelt essay by C. S. Lewis about identity politics.
Surely Lewis would be pastoral, witty, warm, and engaging when interviewed about his essay on ABC radio, making the host laugh. I imagined the delight of discovering a new book by Dorothy Sayers on, well, anything. Dorothy Sayers surely would find a significant group of followers to her Substack, commenting on changing feminist and gender movements or about what it means to be human.
Human or AI error?
At the beginning of each podcast episode, there is a disclaimer that the output is not meant to be authoritative, and that AI can get things wrong like humans. I felt another chill run through me. I certainly make mistakes. Even if I made no discernible errors in spelling or grammar, I might be wrong in my assertions about something far more important. If I can be wrong, then so can Lewis, Chesterton, and Sayers.
Those in the reformed tradition will know that our heroes let us down. Luther, who we celebrate for challenging corruption in the church of his day, was known to make antisemitic comments in his later years. Tolstoy’s late-life conversion led him to almost suicidal despair which may cast shade on some of his work or at least lead us to read with caution.
It’s highly likely that if Tolstoy was writing today, he would shun social media and choose not to allow us to use his ideas. Dorothy Sayers may have chosen silence over blogging about gender issues that are dividing people. Lewis may have avoided many of the culture wars that some we admire engage in. Just because we can take something Sayers once said about femaleness and apply it to our current issues through neural networks, does not mean that she would have chosen that particular approach to the issue.
Recently I was rereading The Great Divorce by C. S. Lewis. This speculative science-fiction novel imagines an afterlife in which the “ghosts” of hell are persuaded to give up their petty jealousies and griefs by “angels” from heaven. Like the creator of the synthetic podcast, Lewis gave a careful disclaimer in the foreword that we were not meant to take his speculative novel as authoritative knowledge about the afterlife.
Read more: ‘A gift we need to steward well’: AI meets Christianity
Yet there is something inspired in Lewis’s speculation about heaven and hell, which draws upon his knowledge of scripture, logic, reason, and relationships. He had personal experience in speaking to people living during a major world war, with death hanging over their heads. Lewis often responded to critics of his work with thoughtful modifications when he believed he was wrong. He humbly admitted that he had access to a finite source of information and that, like other humans, he had unfortunate and unavoidable blind spots.
During this re-read, I took a new pleasure in the moment where the protagonist (ostensibly Lewis himself) meets his hero George McDonald (yet another fiction and nonfiction author) during a bus trip from the gloomy streets of hell into the bright hinterlands of heaven. The imaginary conversation between Lewis and his intellectual hero is fascinating and believable. McDonald helps Lewis understand what is happening, providing commentary and guidance on his experiences in his characteristic Scottish burr.
Lewis used his imagination to produce advice from his heroes after they passed, so perhaps he would not mind becoming a hologram of himself, a light that helps us see this chaotic and disturbing world more clearly. He of all people knew we saw through a “glass darkly.”
In his foreword to The Great Divorce, Lewis warned his readers that his speculative work cannot be taken as “truth.” As AI content becomes more and more believable, we must apply the same warning to the things we create with it. We need to remember that it is not just the outputs that are prone to error, but that the material they draw on is incomplete in ways that will only be clear on the other side of eternity.
Dr Amy Isham is a writer, researcher, and podcast host, writing here on behalf of ISCAST–Christianity and Science in Conversation. She is the Melbourne Engagement Lead of City Bible Forum, which runs Third Space, creating spaces for people to explore life’s big questions.
For more faith news, follow The Melbourne Anglican on Facebook, Instagram, or subscribe to our weekly emails.
