Study honestly with AI

Many tools now have AI in them. AI is technology that lets computer systems imitate human learning and problem solving. These tools can produce reports, images, computer code or maths answers.

The Higher Education Authority (HEA) is a national body. It has a legal responsibility to govern and regulate higher education in Ireland. It says every college should publish one clear policy on AI use. That policy should say which uses are permitted, conditional or banned.

Rules can differ between courses and between assignments. So your own course decides. This lesson cannot tell you what yours allows.

Start with your assignment brief. The brief is the written instructions for a task. The HEA says staff should set out AI permissions in it. It also says you must read it carefully. If it is unclear, ask your lecturer (a college teacher).

The HEA published these documents in December 2025. Its document for students says: “Students are subject to institutional regulations governing academic integrity, assessment, and conduct”. So students are still bound by their college’s own rules.

The same document says students “must not assume that practices permitted in one learning context automatically apply in another”. It says it is “the student’s responsibility to read assessment briefs carefully, to seek clarification where expectations are unclear, and to comply with the specific requirements set by individual teaching staff”.

One university’s example of this, UCL in London, says that further requirements may be set by “a department, academic programme or individual teaching staff, or for a particular assignment”. UCL last updated that page on 29 July 2026.

Quality and Qualifications Ireland (QQI) is responsible for the outside checking of quality in further and higher education in Ireland. In November 2019, it set up the National Academic Integrity Network (NAIN). In 2023, NAIN wrote guidance for college teachers.

Academic integrity means honest and moral behaviour in an academic setting. QQI says it assumes a learner deals honestly with their college, including in the assessments they submit. An assessment is work that a course uses to check what you have learned.

NAIN names a case where integrity “is breached”. It involves GenAI, which means generative AI. That is AI that makes new content, such as text or images.

Students hand in what GenAI made “as their own work without acknowledgement or without authorisation”. That gives two questions. Were you allowed to use AI for this task? And did you say that you used it?

NAIN also says AI tools can give unreliable answers while seeming confident. A guide from University College London (UCL) says AI “may fabricate quotations and citations”. So check what you use. Lesson 1.6, Why they make things up, covers this.

The full NAIN sentence reads: “Academic Integrity is breached if students submit the products of GenAI as their own work without acknowledgement or without authorisation to use GenAI in fulfilling the task.” NAIN’s foreword is dated July 2023, and QQI hosts the file. QQI’s December 2025 lexicon, a list of common terms for the national network, lists “unauthorised use of artificial intelligence including generative artificial intelligence (GenAI)” among its examples of academic misconduct.

NAIN says credit “is awarded for their work not that of others or produced by GenAI systems or obtained from other sources”. The HEA’s principle on academic integrity, transparency and accountability says “the responsibility for academic work must remain traceable to human agents who can defend, contextualise, and take ownership of intellectual contributions”.

Two more lines show where the weight falls. The HEA’s document for students says students “retain full responsibility for all materials submitted under their name, regardless of whether generative AI tools have been used in their production”. QQI’s lexicon, in its entry on fabrication (making up data or other information), adds: “It is the responsibility of the learner to ensure that all material submitted is genuine.”

NAIN’s case above is about what students hand in. So we sort study uses by one question. Does anything the AI made end up in the work you hand in? This sorting is ours. NAIN does not use it.

The activity Help or hand-in? shows eight study uses, one at a time. For each, you predict one of four answers. A use can be fine, fine if you declare it, not allowed, or depend on your course rules.

Then you see our reading and why. Two uses also show a line you could write.

Our readings use only the sources named in this lesson. They are readings, not rulings, and your course’s rule wins. The same tool can suit one task and not another. NAIN says assessment is meant to give the student “an opportunity to demonstrate achievement of the intended learning outcomes” of the course. What a task is meant to show decides what help fits.

One published framework, the AI Assessment Scale, describes five levels of AI use in assessment, from “No AI” to “AI Exploration”. Its authors say it is meant to help teachers and students talk openly about which AI use is appropriate. We do not use its levels here.

QQI’s lexicon, a list of common terms, says AI “should support rather than replace learning”. Our activity uses that line.

A log is a short note you keep each time you use AI for coursework. UCL’s guidance for its students says you must document your use of AI tools and keep your drafts. It says this lets you show your process.

Our log entry has five parts:

• the date;
• the tool and its version;
• what you asked for;
• what you used from it;
• what you checked or changed.

Here is a made-up entry. StudyBot is not a real tool. “9 October. StudyBot 3. I asked for headings for my essay. I used two. I checked each one against my notes.”

A log is not permission. The HEA asks for a declaration, a note that says you used AI, where AI is used “within permitted bounds”. This lesson also calls it a disclosure line. So check the rules first, then log.

The five parts are our suggestion. They match what the HEA and UCL say a declaration needs. The HEA says a declaration should give “the exact name and version of the AI system used” and “a clear description of how the AI was used”. It should also include “representative prompts or describe the general prompting strategy”, and “the steps taken to validate gen AI outputs”. Prompts are what you typed to the tool. A prompting strategy is how you chose what to type. To validate an output is to check it. UCL adds that extra requirements may include the date the output was made, the output itself, and how it was changed.

The HEA’s AI literacy training document also mentions logs. It says participants should be able to do “maintaining logs that facilitate continuous improvement and incident review”. That sentence does not describe a log kept by a student. So the study log in this lesson is ours, not the HEA’s.

A log does not decide what your course allows. The HEA’s words “within permitted bounds” come first.

A disclosure line is a short note in your work that says how you used AI. The HEA calls it a declaration. If your course gives a form, use that form.

The HEA says a declaration names four things. They are the tool, the purpose, how much the tool did, and your checking steps.

Here is a made-up line. “I used StudyBot 3 on 9 October to give feedback on my draft. I rewrote two paragraphs myself. I checked my changes against my sources.”

A line must be true. The HEA says failing to acknowledge AI use is a breach of academic integrity. A line also cannot make a banned use allowed. NAIN names two conditions: you had permission, and you said you used AI.

The HEA says a declaration should be “specific enough to support verification”. It says one should also give “an honest evaluation of the relative contributions of human and AI to the final work”. It says a declaration belongs in the work itself: for a written assignment, “immediately after the title page or in a designated section of the methodology” (the part that says how the work was done). For code, it says “documentation headers or README files” (a README is a text file that explains the code).

UCL’s template asks for a minimum: the name and version of the system, its publisher, its web address, and a one-sentence description of how it was used. UCL’s example is: “I acknowledge the use of Microsoft Copilot (version GPT-4, Microsoft, https://m365.cloud.microsoft/chat (opens in a new tab)) to summarise my initial notes and to proofread my final draft.” The HEA’s document for students says failure to acknowledge AI use “will be treated in the same manner as plagiarism, fabrication, or other forms of academic misconduct”. QQI’s lexicon says plagiarism happens “when someone uses another person’s words, ideas, or work without giving them credit”. Its entry for plagiarism begins “Presenting work/ideas taken from other sources without proper acknowledgement”. It lists “Use of GenAI without appropriate authorisation and acknowledgement” as an example of plagiarism for work presented for academic credit.