Artificial Intelligence Use Policy
Universidad Católica Digital de Guadalupe (UCA Guadalupe) · Campus IACampus
Promoting entity: Private Association of the faithful Gaudium de Veritate (the joy of truth) (Toledo, Spain), developed by Instituto Santo Tomás de Educación y Familia S.L. (Pozuelo de Alarcón, Madrid, Spain)
Approved by the promoting entity on September 5, 2026; this
version, on September 27, 2026. It is
drafted based on what the platform does; each commitment in this document corresponds to a mechanism that exists in the code, is auditable, and can be verified.Review: annually, and whenever the use or regulation of AI changes or a new degree programme enters service.
1. Why this university has an AI policy
Because it uses artificial intelligence to create, support, and transmit knowledge. This policy exists so that anyone
—student, professor, academic authority, or public authority—can know what AI does here, what it never does, and whom to contact with a complaint.
Three commitments bind us:
- The Magisterium of the Church on artificial intelligence, and in particular the encyclical letter Magnifica Humanitas (Magnificent Humanity) of Pope Leo XIV
(May 15, 2026), on safeguarding the human person in the age
of artificial intelligence. For us, it is not an ornamental reference;
it is the criterion by which what we do is judged. - The principles of the Rome Call for AI Ethics (romecall.org, February 28,
2020): transparency, inclusion, responsibility,
impartiality, reliability, and safety and privacy. We follow its
principles; formal accession is in process. - European and international regulations: in particular, the Artificial Intelligence Regulation
(AI Act) and the General Data Protection Regulation (GDPR).
And one conviction of our own: the person is the end of this university, never the means. A technology that treats the student as an object—of surveillance, measurement, or optimization—is out of place here even if it were legal and even if it worked. People above technology.
Magnifica Humanitas (Magnificent Humanity) and this campus
Leo XIV signed this encyclical on the 135th anniversary of Rerum Novarum (New Things),
and that continuity is not accidental: the question of the machine and work that
Leo XIII opened returns today as the question of the machine and the person. For
a university built on the social doctrine of the Church, that is its
natural place.
From its teaching we retain three things that have a direct technical consequence
on this campus, and are not merely matters of intention:
- Artificial intelligence is not morally neutral. Whoever deploys it
is accountable for it. That is why this policy exists, why every artifact
generated records which model produced it, and why publishing any
content requires authorization from a person. - The machine does not live what it imitates. It can reproduce language and
processes of thought, but it does not love, suffer, possess moral
consciousness, or know from within friendship, work, or responsibility.
Hence, here no AI grades anything that counts toward a degree,
none decides on its own what is published, and none acts as the student’s
spiritual companion: it is listened to and the student is referred to a person. - The educational alliance for the digital age. The school and the
university are not spectators of technical change but the place where
people are educated to live through it. We teach with artificial
intelligence and teach about it, according to the same criterion.
Official text: vatican.va, Magnifica Humanitas (Magnificent Humanity) (May 15, 2026).
The six principles of the Rome Call, and where they appear in the code
UCA Guadalupe commits to endorsing and implementing the six principles of the Rome Call:
| Principle | How it is fulfilled here |
|---|---|
| Transparency | The entire community is informed about the use and scope of artificial intelligence on the Campus. Every AI that interacts with students is labeled as such, and the assistant must declare itself artificial if asked. Each correction, each piece of content, and each response records which provider and model generated it, and on what date. |
| Inclusion | The entire campus operates in Spanish, English, and Chinese, as well as other languages, with translation performed by the platform itself so that language does not determine or limit who can study. In addition, the entire platform is "responsive" so that it can adapt as much as possible to the student. |
| Responsibility | Human approval for what matters. No publication occurs without a person approving it, and who that person was is recorded. Issuing a degree requires two different people: the person who approves it and the person who issues it. |
| Impartiality | Automatic grading is formative and does not count toward the degree. Any student may challenge a correction from the same screen where it is received, and a person responds in writing within a committed time frame. |
| Reliability | No content reaches the student without passing a quality threshold and review by a coordinator; statements concerning faith, morality, or doctrine are also flagged for theological review by a person. The prompts expressly prohibit fabricating quotations and sources, and the use of two adversarial AI models improves the reliability of the proposal. |
| Safety and privacy | Conversation history is retained for a limited time and purged automatically. Records that the law requires to be retained are stored separately and do not contain the student’s conversation. The platform complies with the most demanding privacy and security standards. |
2. What we use AI for
| Use | What it does | Who decides |
|---|---|---|
| Writing content | Drafts lesson material based on material provided by the faculty | A professor or coordinator approves it before publication. Without that approval, nothing reaches the student |
| Grading exercises | Evaluates responses and provides feedback | It is formative: it guides the student and does not decide the final grade |
| Proposing learning paths | Suggests a learning route based on what the student is seeking | The student accepts, changes, or rejects it. The core curriculum is imposed by the university, not by the model |
| Providing support | Campus assistant, lesson assistant, Master of Disputatio (disputation), TFG tutor, and, in each session, assistance with exercises, the reply to your response in the quaestio (question), and commentary on your synthesis | They provide guidance; they may propose grades for a human to assign |
| Providing information to prospective students | The website admissions assistant answers questions about what is studied, how much it costs, and how to apply | It does not admit, charge, or grant anything: it directs users to the appropriate screen. It does not retain the conversation |
| Presenting the lesson with image and voice | Video using the image and voice of a real instructor, generated by AI, or of a synthetic figure | With a real instructor, only if the instructor has given written consent for the use of their image and voice; without that consent in force, it is not generated |
| Translating | Translates content into the campus languages | Reviewable by the faculty |
| Reading aloud | Synthetic voice for the content to improve accessibility | — |
3. What AI never decides
This is the central commitment, and it is not a promise about conduct; this is how
the platform is built.
- No grade that counts is assigned by AI. Automatic grading
is formative and advisory; the instructor assigns the grade. The degree is earned through completed lessons and,
when applicable, through a Final Degree Project reviewed by a person. - No degree is issued automatically. A coordinator approves it and an
administrator issues it: two different people. - No content is published without an instructor approving it.
- No learning path leaves out the university’s common formation: the
server inserts it, regardless of what the model proposes or what the
student prefers.
Legal consequence of the above: no automated individual decisions are made
here of the kind regulated by Article 22 of the GDPR: human intervention is always present in important decisions.
4. What we do not do, and why
We do not infer students’ emotions or states of mind. Neither
frustration, demotivation, nor stress. This is prohibited by Article
5.1.f of the European AI Regulation in educational institutions, and we
are glad that it is: affectively profiling an adult who is studying in order
to act upon them without their knowledge is to treat them as an object.
We care about ensuring that no one gets lost along the way, and we address this in another way:
behavioral signals—days without logging in, repeatedly failed exercises—
with explainable rules that anyone can understand and verify. None of these
estimates how the student feels; they look only at what the student has done, the objective data.
Using those signals, the platform does two things. It may write to the student
an automatic, cordial reminder with no academic consequence whatsoever: it is
a “we miss you; there is still time,” not a warning or a note in the student’s
record. And it alerts a person at the university when it would be appropriate
for someone to take a genuine interest. What it never does is make a decision
affecting the student’s assessment, record, or degree: that is always decided
by a person, as stated in section 3.
The student can deactivate those reminders at any time, and is told how to do so
in the message itself.
We do not use students’ work or conversations to train
models.
We do not profile students for commercial purposes.
5. Transparency: how students know there is an AI
- The platform indicates it, and in the enrolment form the student is informed of the use of AI.
- The campus assistant, lesson assistant, Master of
Disputatio (disputation), and TFG tutor present themselves as artificial
intelligence, and state that they provide guidance but do not grade. - The website admissions assistant warns before the first
question that the response is provided by artificial intelligence. - Videos with an instructor’s image or voice generated by AI indicate this
in the lesson itself. - Exercise corrections carry a notice that they were generated by AI,
are formative, and may be reviewed by the instructor. - The read-aloud function warns that the voice is synthetic.
- Course content does not carry an on-screen label: it undergoes
instructor approval and is published under the responsibility of the
university, like any instructional material. Its provenance is recorded
internally.
6. Data and records
- For every piece of generated content, there is a record of which model produced it and
when, so that it can be audited afterward. - Exercise corrections and content versions are
retained, because they constitute the record of a system for which European
regulations require traceability. - Students’ conversations with the assistants are deleted after
36 months. What someone asks in private says a great deal about that person, and there is no reason
to keep it indefinitely. - Students may exercise their rights of access, rectification, erasure, and
objection by writing to info@iacampus.es.
7. The impact assessment
Before the campus was launched, a fundamental rights impact assessment (FRIA) was conducted, which is what Article 27 of Regulation (EU) 2024/1689 requires of those deploying high-risk AI systems in education. It identifies the affected uses—exercise grading and the proposal of learning paths—the specific risks to students, the measures already present in the code, and those that are still lacking.
It was approved by the promoting entity on September 5, 2026; this
version, on September 27, 2026, and is reviewed every year, and sooner if
the use of AI changes or a new degree programme enters service.
It is an internal working document and is not published: it states by name
what has not yet been resolved, and that is its usefulness. It is made available to
the competent authority that requests it, and to any student who wants to know how what affects them was assessed.
8. Who is responsible
- Ultimate responsibility for the use of AI on the campus rests with the
university, not with the model provider or the person who programs it. - The faculty are responsible for the content they approve, whether it was
written by a person or a machine. - Person responsible for oversight: the President of the
promoting entity is institutionally responsible; operational oversight
—reviewing that what is stated here remains true—belongs to academic
coordination, which also handles complaints. - Review: annually, and whenever the use of AI changes or a new degree
programme enters service.
9. If something goes wrong
Any student may challenge a correction they consider unfair, content they believe to be erroneous, or treatment they consider inappropriate, and their complaint is handled by a person.
The procedure implemented on the campus:
- How: through the link accompanying each automatic correction
(“do you think it is unfair?”), from this same page, or by writing to
info@iacampus.es. - Who resolves it: the program’s academic coordination. Never AI.
- Time frame: a response within 10 business days.
- Record: every complaint and its resolution are recorded, and that
record is reviewed once a year. If several students make the same complaint,
the problem is not the students’.
Approved by the promoting entity on September 5, 2026; this version, on September 27, 2026. Scheduled review: annually.