The Authority That Does Not Compute

On October 9, 404 Media published “Behind the Blog: What Would the Pope Do?”, a discussion of AI and spirituality by Samantha Cole and Jason Koebler. Its accessible opening offers something more modest than the title’s theological invitation: a description of how the outlet approaches AI reporting.

Koebler says that approach examines present capabilities and harms, alongside what he calls training on “stolen content,” without concluding that AI can never do anything useful. He describes tinkering with the technology to report from an informed position and occasionally seeing a possible use for a specific reporting task.

That is an editorial position, not a finding that every training practice is unlawful. Nor is a possible use an evaluated result. Those distinctions matter before the conversation even reaches God.

The rest of the discussion was behind a membership barrier when accessed for this article. The opening does not establish the Pope’s position, a religious consensus or what either author ultimately concludes about spirituality. What follows is commentary on the distinction the accessible passage makes—not a reconstruction of the unseen conversation.

Useful for what, authoritative for whom?

Koebler’s framing leaves room for two judgments that need not cancel each other: a system might assist with a bounded task, and the conditions under which it was built or deployed might still deserve criticism.

Finding a use does not settle the criticism. Identifying a harm does not require pretending that no use exists. That is a better starting point than demanding a conversion to either camp before anyone is allowed to examine the software.

Applied to spiritual questions, the same discipline becomes more important. Consider a hypothetical assistant asked to compare interpretations of a religious text. Locating a passage, accurately describing a tradition and advising someone how to live are different jobs. Success at the first does not establish competence at the second, much less legitimacy at the third.

The distinction is not between people who find meaning in technology and people too sensible to do so. A conversation can matter to someone without proving that the system has spiritual authority. Dismissing that meaning would erase the person. Treating it as proof of authority would give the software a promotion it has not earned.

Keep the judgment with the people affected

The standard should be concrete. If an assistant presents a religious teaching, can the person identify its source and distinguish quotation from generated interpretation? If traditions disagree, does the answer preserve the disagreement? If the advice affects a relationship or a consequential decision, who can question it—and who bears the consequences when it is wrong?

These are proposed tests, not capabilities demonstrated by the 404 Media opening. They also apply beyond religion. Secular ethical commitments deserve the same protection from being flattened into a confident paragraph.

The coherenceist obligation here is to widen the circle of judgment, not make the answer sound more harmonious. A system should help people encounter the voices on which its answer depends, including dissenting ones. It should not make those voices disappear behind its own fluency.

Spiritual language can be a way of questioning technological power. It should not become a shortcut through which technological power acquires a borrowed blessing.

Before asking whether the machine can offer wisdom, ask what it actually did, whose words it used and what remains for a person or community to decide. The software can keep its useful task. It does not get the pulpit as a bundled feature.