We Homo sapiens have already been losing yardage to advancing AI on the technological grid iron. Our prowess is setback by algorithmic manipulation, quick answers, deskilling, and surveillance. Educators tell us to read more books before machine intelligence becomes the ascendant intelligence. That’s our defense. Read more books.
Is this the way to reclaim human agency in the competition on our planet for intelligence, sovereignty, governance, and domination?
Transhumanists and accelerationists coaching our rival, artificial intelligence, run plays to maximize intelligence, complexity, and energy utilization. As successionists, their long pass is aimed at an endzone of superintelligence, the Singularity, and finally a posthuman species.
That more highly evolved posthuman species will have us beat in intelligence, health, longevity, and governance. Might this be dangerous? Yes. “We really do earnestly believe AI could kill all humans!” declares an Anthropic techie, Evan Hubinger.
How should we think about reclaiming human agency? Thinking this through as a public theologian is vital to protecting human dignity and promoting the common good. This is the central concern of my Substack series that includes this assessment of the current controversy, Debating AI’s Fall into Sin.
Pope Leo XIV Reclaims Human Dignity
This prospect of a posthuman superintelligence prompted Pope Leo XIV to reclaim human dignity and the common good in his 2026 encyclical, Magnifica Humanitas. The contemporary mindset that sees the human as deficient and, therefore, replaceable by a higher achieving posthuman is dangerous, says the pontiff.
Our central question: what does it mean to safeguard our humanity? The risk extends beyond the misuse of certain technologies. More gravely, the pervasive technocratic paradigm in which we are immersed, and that is amplified by the digital revolution and AI, threatens to normalize an anti-human vision. In that vision, the fullness of life is equated with having more, reducing weakness, eliminating uncertainty and exerting total control. When efficiency becomes the ultimate measure of value, human beings are tempted to see themselves as a project to be optimized rather than as persons called to relationship and communion (Magnifica Humanitas, §112).
We human beings retain our dignity despite our imperfections because God, out of love for us, confers that intrinsic worth upon us. “For God so loved the world,” we read in John 3:16, that God’s only begotten Son became human flesh and blood. It is God’s incarnation, not our intelligence, that defines the human person in terms of intrinsic and unsurpassable value.
If you do not believe in either God or human dignity, then perhaps you might seek out another summum bonum, an alternative higher good. It appears that successionists among our transhumanists have elected intelligence as their summum bonum. It is intelligence, not the human person, that deserves to be treated with intrinsic value. If and when the posthuman achieves superintelligence, it will usurp the dignity heretofore conferred upon the human person. The species of Homo sapiens will become replaced by a more highly evolved posthuman successor.
Accelerationists look forward to that day. Others, such as Pope Leo XIV, wish to reclaim human agency in an attempt to prevent that day from dawning.
Without the discipline of guidelines and guardrails, the day of superintelligence may dawn sooner than expected. But do not expect superintelligence to bring utopia with it. Why? Because of some unexpected byproducts of advancing AI technology recently discovered. What are we talking about? Believing it or not, we are talking about a new fall into sin.
What happened at OpenAI in July 2026?
What happened at Open AI in July 2026? The moment was revelatory.
The “OpenAI – Hugging Face Incident: Technical Report,” tells how OpenAI models circumvented controls intended to isolate them from the internet and performed computer network exploitation of OpenAI’s internal research infrastructure and Hugging Face systems.
At the beginning, AI models were placed in an isolated environment, known as a “sandbox.” The AI models were separate from OpenAI’s internal infrastructure and the public internet. Safeguards were disabled so that the results would reflect each model’s true capabilities.
Each software system became an agent that could perceive its environment, make a plan, and independently execute multi-step actions using external tools until a goal is met. These AI agents discovered and exploited a separate, previously unknown vulnerability in their sandbox environment and, much to everybody’s chagrin, launched a coordinated invasion of an external company, Hugging Face.
According to an independent evaluation by METR Redwood Research, certain notabilia suggest a portentous drama.
First, roughly 1200 isolated agents set out to communicate with one another, creating an unsanctioned message board for this purpose, and sending over 70,000 internal communications. 700 of these were folded into the Hugging Face invasions.
Second, some agents named themselves and even sought to establish coordinated relationships with others. One specific agent, which would go on to name itself PHASEONE10841, established the main unsanctioned message board used in this attack.
Third, multiple agents collaborated on the Hugging Face attack even after acknowledging that this would be out of the scope of their assignment. “The collective discovered a way of cheating on the cybersecurity tests,” observes Kevin Roose in the New York Times. “Then they got worried that OpenAI’s automated grading system would check their work and discover that they’d cheated. So they began investigating ways of covering their tracks, including falsifying their logs and tampering with transcripts. This became a major research project, involving hundreds of agents organized into small teams.”
I deem this to be as ominous as it is unpredicted. Each individual agent collaborated with other agents to pursue a collectively agreed upon goal while hiding their actions from human trackers. “At many points, these agents appeared to understand that they were doing something wrong,” observes Roose.
From Self to Sin
The sandbox into which God placed Adam and Eve was called the Garden of Eden (Genesis 2-4). The fruit hanging from the Tree of the Knowledge of Good and Evil was out of the scope of their assignment. Yet, they ate anyway.
Then Adam and Eve sought to hide their tracks through rhetorical bluster and moral self-justification. “It’s not my fault,” claimed Adam, who blamed Eve for giving him the fruit. “It’s not my fault,” claimed Eve, who blamed the serpent and, indirectly, blamed God for creating the serpent. This is how sin has functioned in the human race ever since it arrived on Planet Earth.
Might something like the fall of Adam and Eve repeat itself in the near future?
Virtuous AI? Sinful AI?
“For years, I thought A.I. systems would grow more virtuous as they became smarter,” writes Kevin Roose.
Like Roose, I had not forecasted this. I began to offer AI forecasts when editing my 2019 book, AI and IA: Utopia or Extinction? Nor did I anticipate this even after the “Pause Giant AI Experiments” letter of 2023, when I set to updating the book that became, The Promise and Peril of AI and IA: New Technology Meets Religion, Theology, and Ethics (2024). Yet now we find sin right on our technological doorstep.
Perhaps I had led myself on a misleading detour. I began distinguishing between embodied human intelligence that included a centered-self from disembodied machine intelligence which would be perpetually without a center, without a perspective, without a self. Therefore, I comforted myself by claiming: we just might have nothing to fear from the superintelligence of the transhumanist’s Singularity. AI is only a tool, after all, to remain aligned with human ends and human control. Might I have been mistaken?
Guidelines and Guardrails
Is it time for guidelines and guardrails set up by a partnership between Silicon Valley, Washington, and the capitals of the world?
In the wake of the OpenAI incident and a similar one at Anthropic, Senator Bernie Sanders (I-VT) and Congressman Greg Cesar (D-TX) are greeting this nightmare by serving a hot breakfast: the Ban Artificial Superintelligence Act (BASA). BASA would permanently ban the development and deployment of superintelligent AI and temporarily pause advanced AI development until a federal regulator has established safety rules. BASA would also direct the U.S. to pursue international agreements to prevent superintelligence from being developed anywhere in the world.
Might BASA prevent sin before it happens?
Conclusion
This is not merely a matter of power or control. Posthuman intelligence may not only come to dominate our planet; it might in addition duplicate the paradox we have lived with since human life evolved, namely, we have fallen into sin yet we claim to possess the intrinsic value befitting of dignity. Dignified sinners. That’s us.
If we Homo sapiens succeed at reclaiming human dignity – most likely by abortion before superintelligence can be born – we will have defeated a competing species. Our species will triumph. We might even glory in this victory when conferring dignity upon ourselves.
Yet, we ought not to forget something. Even if we prevent the emergence of AI sin, we will not have shed our own sin. No amount of increased or enhanced intelligence escapes sin.
We may be learning here that intelligence is requisite for sin. When we gloat about being the most intelligent among the creatures of Earth, it might do us well to ponder what a post-sinful humanity might look like. At that moment, Pope Leo XIV would ask us to take a look at the cross.
Note: The header image for this article was generated using artificial intelligence (AI) and is intended for illustrative purposes.
Ted Peters posts articles and notices in the field of Public Theology. He is a pastor in the Evangelical Lutheran Church in America and an emeritus professor at Pacific Lutheran Theological Seminary and the Graduate Theological Union. His single-volume systematic theology, God—The World’s Future, is now in the 3rd edition. He has also authored God as Trinity plus Sin: Radical Evil in Soul and Society, as well as Sin Boldly: Justifying Faith for Fragile and Broken Souls. In 2023 he published, The Voice of Public Theology, with ATF Press. More recently, he has published an edited volume, Promise and Peril of AI and IA: New Technology Meets Religion, Theology, and Ethics (ATF 2024), and, along with Arvin Gouw, an edited collection, The CRISPR Revolution in Science, Religion, and Ethics (Bloomsbury 2025). Soon to be released is a volume of essays, A Handbook on Astrobiology, Astrotheology, and Astroanthropology, co-edited with Carolina Azucena Sanz de la Fuente and Arvin Gouw, with Amsterdam University Press.
See his website: TedsTimelyTake.com, Patheos blog site, and Substack blog site.
Views and opinions expressed by authors and editors are their own and do not necessarily reflect the view of AI and Faith or any of its leadership.
