AI agreements at school and university: what students and parents really have to sign in 2026

StudierAI · 22 Jun 2026 · 9 min read · Genitori

In 2026, many schools and universities are introducing (or updating) an “AI Pact” or a policy on the use of artificial intelligence. For parents it may seem like yet another document to sign, but in reality it’s something very concrete: it sets out what is allowed with tools such as chatbots, text generators, summarizers, and study assistants, and what is instead considered improper during homework, assignments, and exams.

In this article you’ll find a practical guide, based on practices that are becoming standard in Europe and in Italy: how to read the clauses, how to understand what “academic integrity ai” really means, what to expect from “university proctoring 2026” and from “cheating ai detection exams” systems, and above all how to help your child use AI to study in a lawful and transparent way.

Why in 2026 they ask you to sign an “AI Pact” (and what changes for families)

The main reason is simple: generative AI has become a “study tool” as widespread as the calculator or the spellchecker, but with a much broader impact. Schools and universities are therefore formalizing rules that used to be implicit, to avoid uncertainty and different treatment across teachers and courses.

In 2026, moreover, two “system-level” forces come fully into effect: on the one hand the European framework (AI Act) and the growing focus on data protection; on the other hand internal quality and assessment policies (transparency of tests, traceability of submissions, consistency between learning objectives and exam formats). In practice, the “school artificial intelligence policy” and university codes of conduct are trying to answer one question: how do you assess real skills when tools exist that can produce text, code, and plausible solutions in a few seconds?

For families, above all one thing changes: the signature is not just “acknowledgement.” It is often an extension of the shared responsibility agreement (in school) or the university regulations (at university). It covers:

teaching activities in class and at home (homework, reports, papers);
platforms and accounts (Google/Microsoft, LMS, institutional email, submission tools);
written and oral assessments, including online formats and remote sessions;
use of external tools, including “off campus ai” services (i.e., not provided by the institution) and apps installed on personal devices.

The reassuring point is that these documents, when well written, are not meant to “ban everything.” They are meant to create boundaries: what you can do to learn better and what instead alters assessment. Reading them carefully avoids surprises (for example, penalties for failing to disclose AI use) and helps set healthy habits from the start.

The typical clauses: what is allowed, what is “cheating,” and what must be declared

The most useful pacts and policies distinguish between **lawful ai study use** (support for learning) and improper use (replacement of the assessed performance). The recurring clauses can be “translated” like this.

1) **What is allowed** (almost always): using AI as a tutor. Practical examples:

asking for alternative explanations of a concept, with examples and counterexamples;
creating review quizzes and flashcards from one’s own notes;
running oral-exam/interview simulations, with progressively harder questions;
getting feedback on clarity, structure, and grammar of a text written by the student (without generating new content).

2) **What is “cheating”** (almost always): submitting as one’s own an AI-generated assignment or using AI during an unauthorized test. This is where the concept of **academic integrity ai** comes in: integrity is not only about “copying,” but also about disguising the origin of the work. Typical examples of violations:

having AI write an essay/report and “cleaning it up” with paraphrasing to make it less recognizable;
using a chatbot to answer questions in an online test when individual work without aids is required;
generating code or solved exercises and submitting them without understanding and without disclosure, especially in courses where the solution is what is being assessed;
using “off campus ai” tools during a proctored exam or in the classroom when access to devices and external services is explicitly forbidden.

3) **What must be declared (disclosure)**: many institutions ask you to indicate whether and how AI was used. This is not a detail: often the penalty arises not from the use itself, but from the lack of transparency. The most common forms of disclosure are:

a note at the end of the assignment (“I used an AI assistant to…”)
an attachment with the main prompts and relevant outputs, especially for longer work
selecting a checkbox on the platform (“AI used: yes/no”) with a brief description.

A useful example to make it clear: if a student uses an off campus ai tool to turn their own notes into 20 review questions, it is usually allowed. If instead they ask the tool to write the entire assignment and then submit it, that is almost always considered “cheating,” even if the text is reworked. The guiding criterion is: **is AI helping learning, or is it replacing the assessed performance?**

Proctoring, AI detection, and privacy: what data they can collect and what real risks exist

When people talk about “university proctoring 2026,” they generally mean the monitoring of online (or hybrid) exams via software that checks identity, environment, and behavior during the test. The most widespread systems combine: ID/face verification, browser lockdown, webcam and microphone recording, monitoring of active windows and, in some cases, automated analysis of movements and sounds.

Alongside proctoring, some institutions use “cheating ai detection exams” tools or automated detection of generated text (AI detection) for homework and assignments. Here it’s important to be realistic: detection is not a “truth machine.” Detection models can produce **false positives** (human texts flagged as AI) and **false negatives** (AI texts not flagged), especially when:

the student writes in a very “standard” or repetitive way (typical of someone who is learning);
sign up for free
who we are
the student used lawful tools (proofreader, translator, style suggestions) that “standardize” writing.

In short: in 2026 signing an AI Pact does not mean accepting “limitless monitoring” or banning technology. It means defining clear rules to protect assessment and privacy. The safest approach for families is to focus on transparency, traceability, and dialogue: ask what is allowed, disclose when required, and use AI mainly for review and understanding, not to replace personal work.

On the privacy front: a well-made AI pact refers to a (GDPR) privacy notice that explains what data are processed, for how long, on what legal bases, and who the providers are. In proctoring, data may include images, audio, system logs, and session metadata. In the case of off campus ai tools, the main risk is different: entering into prompts personal data or protected materials (non-public assignments, exam prompts, third-party data) that end up on external services not controlled by the school.

Practical questions to ask (or to help your child ask) before accepting proctoring or detection tools:

What data are recorded (video, audio, screen, logs)? How long are they kept?
Who is the provider? Do the data stay in the EU or are they transferred outside the EU?
Are there equivalent alternatives if the student cannot use that system (technical reasons, accessibility, privacy)?
If an assignment is flagged by detection, what is the procedure? Are additional proofs required (drafts, interview)?

These questions are not “mistrust”: they are part of normal due diligence, like reading a privacy notice or asking for clarification on a policy. In many schools and universities, administrative offices and teachers are already used to answering and appreciate a collaborative approach.

How to use AI to study without violating policies: a checklist for students and parents

Come usare l’AI per studiare senza violare i regolamenti: una checklist per studenti e genitori

The most effective (and simplest) strategy is to treat AI as a support that can be documented. If your child is unsure, the practical rule is: **ask first, always disclose when required, and keep records**. Here is an operational checklist, designed for homework, reports, and exam preparation.

**Pre-assignment / pre-exam checklist**

Read the instructions: does it say “AI allowed/forbidden”? If it’s not clear, ask the teacher for a brief confirmation.
If AI is allowed, define the use: tutoring (explanations, quizzes) vs production (final text). Keep production under human control.
Do not enter into prompts sensitive personal data, non-public exam prompts, or third-party information. It’s good practice even when use is allowed.
Keep records: main prompts, document versions (drafts), notes, and sources consulted. In case of a dispute, they help more than any “verbal explanation.”
Disclose use when required: one clear line on what was done (e.g., “I used AI to generate review questions and for grammar revision”).

Two “ready-made” sentences that often resolve doubts in a collaborative way:

• “For this assignment, can I use AI only for review and revision, but not to write parts of the text? Would you like me to declare it in a final note?”

• “If I use an off campus ai tool to create quizzes from my notes, is that okay? Is there a disclosure format you prefer?”

Finally, one piece of advice that really works: train your child to “demonstrate understanding” beyond the submitted text. For example, prepare a brief oral explanation of the steps, or a hand-drawn concept map. This reduces the risk of misunderstandings with detection and strengthens learning, regardless of the specific rules.

Where StudierAI can help (in a way compatible with academic integrity)

Dove StudierAI può aiutare (in modo compatibile con l’academic integrity)

Tools likeStudierAIcan be useful if set up with a **assisted study** logic: helping to organize, review, check understanding, and prepare for oral exams, without “replacing” the assignment that will be assessed. This approach aligns with many academic integrity ai policies, because it makes AI a support for the process and not a generator of the final product.

Here are some uses typically compatible with AI pacts (bearing in mind that the rule is always the teacher’s or the university’s):

**Guided summaries**: starting from one’s own notes and getting a structured synthesis to verify and correct, keeping the student responsible for the content.
**Flashcards and quizzes**: turning chapters and notes into questions of increasing difficulty, useful to avoid “passive” studying.
**Oral simulations**: practicing answering, asking for follow-up questions and clarifications, as a tutor would.
**Planner and organization**: planning study sessions, weekly goals, and reviews, reducing stress and procrastination.

To stay within the boundaries of the policy, the best practice is to make AI use “auditable”: save important prompts and results, note what was used for review and what for revision, and keep a personal version of the work. If you want to explore an assisted-study approach, you canstart for freeorsign up for freeand see which modes best fit the rules of your school or your course. If you’re interested in understanding the setup and principles, you’ll find more details on the pagewho we are.