Toward the very end of Tom Stoppard’s absurdist play, Rosencrantz and Guildenstern are Dead, one of the titular characters (we’ll call him Guildenstern — though the script plays on their interchangeability), reflects, “There must have been a moment, at the beginning, where we could have said — no. But somehow, we missed it.” Perhaps because the PCUSA’s General Assemblies can, at times, remind me of absurdist plays, the line came to mind as I was engaged in the various processes, debates, and actions of the recent 226th General Assembly, especially with regard to matters related to artificial intelligence.
To its credit, the , with a particular emphasis on the destabilizing potential AI has for employment, the economy, civil governance, and global conflict, increasing existing technological disparity, and the theological implications for its usage in the work of the church.” That resolution also highlighted several existing policies that might offer insight into the use of technology more broadly as well as pointing towards a values framework for the use of AI that was developed in association with the United Nations and could be useful in offering guidance to the denomination and its various bodies between now and the time this new policy is approved. (Full confession: I am a member of the Advisory Committee for Social Witness Policy, which originated this resolution, and I played a role in developing it — though I did not write the rather ungainly sentence quoted above). The study team responsible for creating this new social witness policy on AI will include members of ACSWP as well as others with expertise in the field and will bring the report to the 227th General Assembly in 2026 for review and approval.
The resolution also directs the denomination’s Research Services to “conduct a study to gauge the perception of the wider church on artificial intelligence, with a particular emphasis on the internal workings of all councils of the church,” which will be a resource for the team writing the denomination’s new policy. This makes sense: AI currently sits in the awkward spot of being increasingly used by a tech-dependent church even as perceptions of AI tend towards the apocalyptic in popular imagination. Having a sense of both what the church is doing with AI and what the church is thinking about AI will be valuable for the team drafting policy. “Know thy audience” is the first principle of good writing, after all.
The catch, obviously, is that AI is advancing so rapidly that either a new policy could be outdated by the time it is approved or a new policy could be so broad and theoretical in order to still be pertinent that, ironically, it would not be especially applicable in practice. The study team may be caught between the Scylla of understanding rapid technological advancement and the Charybdis of shaping durable social policy.
Because AI technology is advancing fast. AI systems are becoming more sophisticated quickly (ChatGPT 5 is expected in early 2025). And the number of companies developing generative AI is expanding rapidly (my AI-assisted Google search said that there were nearly 15,000 AI start-ups as of over a year ago — and these are only the startups that we know about). Whatever else generative AI is doing, it is generating business and doing so at an exponential rate.
One piece of evidence for its generative power is that even as the PCUSA took up the matter of Artificial Intelligence at a General Assembly, it was using AI in the General Assembly. Simultaneous translation during Zoom meetings, for instance, was AI-driven because that’s now how Zoom works — and since the pandemic, Zoom-use is S.O.P. throughout the church. As General Assembly revealed, the church’s use of AI is outpacing its considered reflections about such use.
Such may be the way of the church’s encounters with new technology. Pope Eugene IV didn’t send a small band of clerics to Mainz, Germany to have a conversation with Johannes Gutenberg about his new moveable type printing press in 1440. Nor—so far as I can determine — did Philo Farnsworth meet with the leaders of the Church of Jesus Christ of Latter-day Saints (of which he was a member) or, for that matter, with Pope Pius XI or any other church leaders, after he invented the television in 1927. Yet it is hard to underestimate the shaping power of either the printing press or the television for the church and wider societies. Both have been great boons for and wreaked havoc in the church. At least to judge from its history, the church’s narrative about new technology is less about it being Prometheus, and accepting the risks of bringing new god-like powers to the people, and more like children playing with toys and trusting that, by the providence of God, those toys won’t be fatal.
The good news is that even as the church uses AI prior to understanding AI (which understanding is in short supply even among its creators), the types of generative AI that it uses — large language models — are a long way from the types of dangers that may come with artificial general intelligence. The less good news is that (a) the use of generative AI still carries costs that are seldom considered by those using it; and (b) such use may tend to inure its users to its current and potential costs. Contemporary AI’s tendency to replicate patterns of misogyny and racism will hardly stir up much outrage in any culture that has become used to misogyny and racism. Contemporary AI’s voracious use of rare resources and energy won’t much affect people glued to their screens and comfortable with their cars. Contemporary AI’s tendency to produce impersonal messages and flat prose probably won’t feel especially costly in any society that treats SMS language/ textspeak (a.k.a. txtspk) as thoughtful communication or reality television as worth its watching. The real dangers of AI aren’t nearly so sexy as the imagined ones.
Which is to say that the costs of AI may be of the mundane sort about which the church has a habit of shrugging its collective shoulders even as such mundane costs, compiled, lead it to places it does not understand, pursuing practices for which it is ill-equipped, and based on perceptions that are blinkered at best. Hopefully, the 226th General Assembly has initiated important steps towards giving AI the kinds of careful, considered reflection that will help the church engage those places, practices, and perceptions. Realistically, such careful and considered reflections will mostly not reach the wider church and the vast majority of its members, even as they become increasingly comfortable with AI (how often do any of the church’s social witness policy statements actually reach congregations, after all?). Yet the church has always been fortunate to live between hope and realism because it is, ultimately, sustained not by its technologies, nor by its considered reflections on them, but by the grace of a God who is the giver of all good gifts and whose purposes are worked out in, with, and beyond the church and its technologies. Perhaps there was a moment, at the beginning, when we could have said – no. Missing that moment, though, doesn’t undo God’s far more profound yes. Perhaps the actions of the latest PCUSA General Assembly with regard to artificial intelligence can be read as revealing that.
