Professional Documents
Culture Documents
NAIN Generative AI Guidelines For Educators 2023
NAIN Generative AI Guidelines For Educators 2023
Generative
Artificial
Intelligence:
Guidelines for
Educators
2 Generative Artificial Intelligence: Guidelines for Educators
CONTENTS
Introduction 2
Part 1: Recommendations 2
What everyone needs to know 2
Introduction
4 Generative Artificial Intelligence: Guidelines for Educators
Introduction
These Generative Artificial Intelligence Part 2 provides the detailed guidelines for each
(GenAI) Guidelines for Educators have been of the four headings.
developed by a National Academic Integrity
Part 3 provides a list of further reading, links,
Network (NAIN) Working Group, as a response
and resources.
to the swiftly evolving and developing field
of GenAI and the challenges that it may pose These Guidelines are not intended to replicate
for academic integrity to educators and their what is currently available to educators, but
students. to provide practical advice which can be
applied to the design of programmes including
They are designed to help support educators
assessments, and be incorporated into
in their understanding of the potential uses of
teaching practice. This area is a dynamic one
Generative AI, both in supporting learning for
which is developing at a very fast pace; this
their students, and most critically, in providing
means that these Guidelines, which will be
a potential ‘short cut’ to students in the
published online, may well need to be regularly
fulfilment of required tasks for assessment.
reviewed and updated.
The guidelines have been written to provide
support and advice for educators to reflect on, We hope however that they will provide some
and as appropriate, to share and discuss with practical support for everyone involved in
their students to enable them to understand tertiary education and that they will have a
and appreciate what is and isn’t permitted. The positive impact in supporting high quality
overall goal is to enable an understanding of teaching, learning and assessment policies
what GenAI can and can’t offer, and thereby, and practices.
to ensure an ethical basis for the use of GenAI
tools, helping students to build their own self-
awareness and knowledge, and avoid breaches Billy Kelly
of academic integrity. Chair, National Academic Integrity Network,
July 2023
The Guidelines are structured so that Part 1
contains a list of Recommendations, classified
under the four headings:
01
Recommendations
8 Generative Artificial Intelligence: Guidelines for Educators
Part 1: Recommendations
02
Detailed
Guidelines
14 Generative Artificial Intelligence: Guidelines for Educators
internationally to consider potential forms (2) Increasingly, educators are sharing use
of regulation, although quite what shape cases of where AI tools might be useful
these will take is very uncertain. in designing courses, lesson planning,
assessment design, student feedback, etc.
(4) As technologies develop, appropriate roles
Whilst as the technology improves, there
in education may be found, but if they
is no doubt that GenAI will be more useful
are used to bypass rather than support
in many professions and contexts, we still
thinking, or used to acquire academic
urge caution, particularly in terms of the
credit which has not been earned via real,
accuracy of its outputs and the fact that
intellectual engagement with the subject
it is largely based on statistical models
of study, then they can undermine the
and hence reproduces or mimics rather
educational enterprise. Part of the current
than being capable of generating original
challenge for educators, learners, and
content. There are many cases where a
institutions is to determine the range of
traditional search strategy (using a search
beneficial applications which might be
engine or library tools) is much more
afforded by such tools.
effective and has the added benefit of
identifying original sources and allowing
you to make judgements about accuracy
What lecturers and other
and quality as well as to provide due
educators need to know and can
credit to the original authors. This is a key
do
point that should also be included in any
(1) Ensure that you are aware of recent and training of students about information
new Generative AI tools, including those gathering and research.
associated with supporting writing, and
(3) If you are using GenAI tools in your own
others which might be relevant to your
academic practice, then you should
particular discipline, such as those which
model appropriate use with your students,
assist the development of computer code,
acknowledge where GenAI has assisted in
solve mathematical problems, generate
your work, etc. It is important not to input
graphics, video, and audio content,
personal, private, sensitive, or copyrighted
etc. Try to understand what these tools
materials (including student work) into
can and can’t do (including by trying
such tools unless you have appropriate
them out on your own assignments or
levels of legal guarantees. Many of the
subject content), and, in particular, avoid
freely available tools at the moment do
mistakenly attributing subject expertise,
not have such protections in place.
intelligence, or capacity for reflection to
such software. It is useful also to be aware
of ethical and legal concerns over how
such tools are developed and used, as well
as the potential inequality of differential
access (e.g., on the basis of cost).
16 Generative Artificial Intelligence: Guidelines for Educators
(4) In designing assessment, it is important (6) There will be cases where it is clearly
to ensure validity, i.e., that assessment inappropriate to use GenAI and, in those
should be a means whereby each student cases, state this clearly in the instructions
demonstrates that they have attained provided to students and ensure that
the intended learning outcomes to an you are able to make the format of the
appropriate standard and that this should assessment sufficiently secure and
be the basis on which academic credit is robust.
awarded.
(7) Academic Integrity is breached if students
• It is important to clarify and reiterate submit the products of GenAI as their
this point with students and to own work without acknowledgement
demonstrate how the assessment fits or without authorisation to use GenAI
this intention. in fulfilling the task. It is important to
ensure that students are informed that
• It may be appropriate to revisit your
this constitutes academic misconduct
original learning outcomes to ensure
and that they should only be seeking
that they are aligned with how you
credit for work which they have
teach, the tasks you set students, and
produced themselves, and that they are
the chosen forms of assessment.
responsible for correctly referencing and
(5) If you intend for students to use GenAI or acknowledging sources and resources
AI-supported tools in any assignment, used in their work.
make this clear. It may not be possible
(8) Review all assessments and assignments
to simply state ‘do not use AI for this
in terms of whether or not they are
assignment’ given the technology’s
susceptible to being completed
ubiquity, so the emphasis should be
successfully by an AI tool without the
on what the student actually does to
student having to engage intellectually,
demonstrate their attainment of the
or personally, with the subject. Some
intended learning outcome with which
common types of assessment should be
the assignment is aligned. If the student
no longer considered to be sufficiently
constructs a report or essay through
robust to award scores which count
clever ‘prompt engineering’ which could
towards official grades. These may still be
have been done by anyone not taking
perfectly valuable for self-assessment and
this module, then this does not provide
formative practice. These include:
the basis for an award of marks or
credit. Where students are expected to • Take-home essays, reports, or similar
use AI tools, ensure there is clarity on documents focused largely on subject
how and why they used such, with an knowledge content and with an ‘all
appropriate declaration on any submitted or nothing’ submission by a single
assignments. deadline and where marks are based
on structure, style, and information;
National Academic Integrity Network 17
(12) Consider what scope there is for content, so, again, be careful about the
assessments which: requirements.
• Focus on ‘process rather than product’. • Are based on whether the learning
In other words, where credit is outcomes have been achieved by the
associated with the various tasks or student, rather than leading to a grade.
stages that the student goes through
• Use in-class writing assignments or
in order to lead to the final product.
problem-solving tasks.
This can be done by breaking down
assignments into key stages and • Are based on, or partially include, an
incorporating student reflection/ oral component in which the students
reporting on how they found their are asked to answer questions around
sources, what aspects they explored, the topic and how they approached the
what challenges they overcame, etc task. There are many examples where
(or whatever is appropriate to the orals have been used successfully,
particular assignment) so that there is even in large class contexts, and in
a developmental journey. which there are particular formats
(interactive orals for authentic
• Are clear in terms of whether the
assessment) which align with the
emphasis is on the development of
discipline or profession. Where it isn’t
writing skills (something the AI tools
practicable for a very large cohort, at
can do very well) or on understanding
least consider undertaking a number
of the specific topic and assess each
of orals either on the basis of random
of these with appropriate methods and
selection or to clarify any concerns
rubrics.
in a particular submission (i.e., this
• Are reflective and based on personal/ is where the main assignment is an
professional experiences and which essay, report, video, etc, and the oral is
are authentic to the discipline or to check on integrity and/or on student
profession. Note that many of the AI understanding).
tools will readily invent ‘reflective’
(13) Work with colleagues to develop an
content if so asked, so try to ensure
effective strategy for assessment which
that the task is genuinely connected
ideally would address issues of student
to the student, or the specifics of
(and staff) workload, the scheduling of
particular aspects discussed in class.
deadlines/due dates, and ensuring that
• Are in different or multiple formats, programme level outcomes are being met.
such as video or audio content, a
• We know that academic misconduct
mix of presentation (with questions
can arise when people inappropriately
and answers) and ancillary materials
react to:
(e.g., reference list, handouts), etc.
AI tools can produce slide decks ◦ ressure and stress caused by too
p
and write scripts for video/audio many overlapping deadlines;
National Academic Integrity Network 19
(6) Refresh study/academic skills supports (2) Ensure that students understand what
for students to provide clarity on is required of them for any assignment
acceptable as well as unacceptable uses or assessment, including the extent to
of such technologies. which there is any allowable use of such
technologies and how this may need to
(7) Work as a sector on issues regarding
be acknowledged. All assignments and
technological developments, data
assessments are designed to determine
protection, protection of IP, and ethics.
whether students can demonstrate
(8) Engage with QQI and other awarding or attainment of the specified learning
accreditation bodies on these issues and outcomes of their module or programmes.
ensure that guidelines and regulations are In other words, the credit is awarded for
clear to all staff and students. their work not that of others or produced
by GenAI systems or obtained from other
(9) Foster an institutional ethos and culture
sources. This means that students need
(reflected in your communications) which
to understand their institution’s academic
emphasises integrity, honesty, trust, and
integrity policy and be able to clearly draw
respect, rather than let the conversations
the line between what is permissible and
be dominated by suspicion, surveillance,
what may constitute misconduct.
and distrust.
(3) Learning is about ‘sense-making’, about
juggling ideas, trying to see where
What students need to know and they fit or where they contradict, about
do rethinking what we thought we already
knew, about seeing things in new ways. All
(1) AI (including so-called GenAI) tools are
of this can be difficult, students can feel
becoming widely available and embedded
vulnerable if something doesn’t appear
in many of the technologies which we use
to be immediately obvious to them and
to write documents, analyse data, design
they need to put in lots of effort or seek
presentations and to support learning. It
help for things to ‘click into place’. It is
is crucially important to understand how
not a weakness or a sign of lack of ability.
such technologies work and be aware of
It’s what university-level education is
their limitations as well as their apparent
about, but it should be counter-balanced
strengths. They are not ‘intelligent’ in
by the sense of achievement and the new
any real sense, nor are they experts on
perspectives and skills that students
topics, but rather largely rely on statistical
ultimately acquire through this effort.
predictions of word combinations or
image features. There are also concerns
about ethical aspects in how such tools
are developed, trained, and deployed and it
is important to be aware of these.
National Academic Integrity Network 21
03
Further
reading,
links, and
resources
24 Generative Artificial Intelligence: Guidelines for Educators
The NAIN GenAI Guidelines for Educators were developed by the Working Group Chair, Iain MacLaren
(University of Galway) and Greg O’Brien (Griffith College), with contributions from Working Group
members - Elva Casey (Hibernia College), Gavin Clinch (ATU, Sligo), Naomi Jackson (CCT College),
Brid Lane (IBAT College) and Cathy Peck (DCU).
National Academic Integrity Network 25
Published by Quality & Qualifications
Ireland (QQI), August 2023 (1st edition)