Our Stance on Artificial Intelligence

The danger is not hypothetical

The last two years have produced well-documented failures of “Catholic answer engine” chatbots: tools that generate confident, fluent, and wrong statements of Church teaching, with no source a reader can check and no human who can be held responsible. Whatever the intentions behind such tools, we think the design itself is the error: a machine should never be placed where a teacher belongs.

Human-directed, human-reviewed

So we have adopted one rule for everything we build: AI is an editorial assistant. It is never an oracle.

• Humans direct. Every use of AI at Lectio Studies executes an editorial standard written by people: our published translation standard, our content standards, our connection vocabulary. The machine follows the standard; it does not set it.
• Humans review. Machine-assisted work passes through verification designed to catch machine failure: independent cross-checking, blind review, and human editors who bear responsibility for every published sentence. Doctrinally sensitive questions are never resolved by a machine, full stop.
• Machines never generate doctrine. When you ask our tools a question, they retrieve and cite (Scripture, the Fathers, the councils, the Catechism) in those sources’ own words, with references you can open. We do not, and will not, ship a chatbot that composes theology.

We show our work, always

Transparency is not a disclaimer on this page; it is built into the product, on principle: you should never have to wonder whether a machine was involved, or take our word for anything.

• Every connection between passages carries its provenance (where the claim came from and how it was verified) and every one can be disputed by any reader with an account. Challenges are reviewed, and anything touching doctrine goes to human review without exception.
• Every text we translate or modernize ships under our published Sacred Translation Standard, with the original displayed alongside, the source edition named, the method disclosed, alternative renderings documented, and a human editor of record on every sentence. You can compare our rendering against the original yourself: that is the point.
• Every work in the Archive states its source, edition, and license. Every technology-assisted step in our pipelines is documented, and this page and our standards will always say plainly where AI is used and where it is not.

If we ever fall short of this, we want to hear it: that is what our dispute and review tools are for.

Why we bother

We could publish faster without any of this. We do it because trust is the entire point of this platform: a reader standing in front of the Tradition deserves to know exactly who (and what) stands between them and the text. Magnifica Humanitas teaches that technology takes on the character of those who use it. The character we intend ours to take on is fidelity, humility, and accountability.